You are on page 1of 294

From polymers to plastics

From polymers to plastics

A.K. van der Vegt

DUP Blue Print

© VSSD 2002

DUP Blue Print is an imprint of:

Delft University Press
P.O. Box 98, 2600 MG Delft, The Netherlands
tel. +31 15 27 85678, telefax +31 15 27 85706, e-mail

Published on behalf of:

Vereniging voor Studie- en Studentenbelangen te Delft
Poortlandplein 6, 2628 BM Delft, The Netherlands
tel. +31 15 27 82124, telefax +31 15 27 87585, e-mail:
URL on this book:
A collection of digital pictures, printed in the book, can be made available for
lecturers who adopt this book. Please send an e-mail to

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photo-copying, recording, or otherwise, without the prior written
permission of the publisher.

Printed in The Netherlands.

Internet edition
NUGI 831
Key word: polymers


The two words in the title of this book, “polymers” and “plastics”, could be
considered as referring to two different worlds.
In the world of polymers, the properties of chain molecules are in the focus of
attention, and are subject of thorough theoretical studies.
In the world of plastics, the end-use performance of the technically used materials
counts, as well as their behaviour in the various processing operations in which they
are transformed into finished articles.
Nevertheless, these two worlds are closely related to each other. The typical
behaviour of plastics materials, strongly deviating from other materials, can only be
understood on the basis of the chain properties.
In this book an attempt has been made to give a survey of polymer properties, and of
the way these are, on the one hand, governed by their molecular structure, and are,
on the other hand, responsible for the technological behaviour of plastics materials.
As a result of this intention, cross-references are given throughout the whole book:
every aspect of polymer science and of plastics technology is closely related to
practically every other aspect!
This book originates from several series of lectures at the Delft University of
Technology during the years 1978 - 1988, later on integrated into a number of post-
graduate and other courses. Since only a first introduction is given, more detailed
treatment and more scientific depth have been sacrificed to the striving for survey
and integration.

Delft, May 1999

A.K. van der Vegt



1.1. Origin of plastics / polymers 11
1.2. Main categories 14
1.3. The most important plastics 15
1.3.1. Thermoplastics 15
1.3.2. Thermosets 18
1.3.3. Synthetic elastomers 19
1.3.4. Composite plastics 20

2.1. Chain structure 22
2.1.1. Main chain 22
2.1.2. Side groups 25
2.1.3. Copolymers 26
2.2. Chain length and distribution 26
2.2.1. Averages 26
2.2.2. Example of a chain length distribution 31
2.2.3. Measuring methods 33
2.3. Chain regularity 38
2.4. Chain conformations 41
2.5. Chain flexibility 47
2.6. Chain interactions 48
2.7. Cross-linking 49


3.1. Glassy state 52
3.2. Molecular picture 54
3.3. Thermodynamics of the glass-rubber transition 56
3.4. Factors governing Tg 59
3.4.1. Chain flexibility 60
3.4.2. Chain interactions 61
3.4.3. Chain length 62
3.4.4. Effect of time scale 62
3.5. Glass-rubber transition of blends 63
3.6. Methods of measuring Tg 64
8 From polymers to plastics

4.1. Conditions for crystallization 65
4.2. The melting point 67
4.3. The crystallization process 72
4.3.1. Nucleus formation and crystal growth 72
4.3.2. Isothermal crystallization 75
4.4. Crystalline structure 78
4.4.1. Crystalline fraction 78
4.4.2. Crystal morphology 79
4.5. Effect on properties 81
4.6. Liquid-crystalline polymers 84


5.1. Rubbery phase 86
5.2. Transition of rubber to fluid 91
5.3. Fluid condition 92
5.3.1. Viscosity 92
5.3.2. Non-Newtonian behaviour 93
5.3.3. Melt elasticity 97
5.4. Some consequences for processing 97

6.1. Models 102
6.2. Dynamic mechanial behaviour 110
6.3. Integration 112


7.1. Stress-strain diagram 117
7.2. Stiffness and creep 118
7.2.1. Modulus of elasticity 118
7.2.2. Creep 120
7.2.3. Non-linearity 122
7.2.4. Physical ageing 124
7.3. Damping 126
7.4. Strength 130
7.4.1. Tensile strength 130
7.4.2. Long term strength 131
7.4.3. Impact strength 136
7.5. Surface properties 138
7.5.1. Hardness 138
7.5.2. Friction 139
7.5.3. Abrasion 142
Contents 9


8.1. Thermal properties 144
8.1.1. Brittleness temperature 144
8.1.2. Softening 145
8.1.3. Thermal expansion 147
8.1.4. Thermal conduction 148
8.1.5. Maximum temperature of use 150
8.1.6. Burning behaviour 150
8.2. Electrical properties 151
8.2.1. Electric resistance 151
8.2.2. Dielectric properties 154
8.2.3. Electric strength 155
8.3. Optical properties 155
8.4. Effects of environment 156
8.5. Stress corrosion 158
8.6. Diffusion and permeability 159


9.1. Polymer blends 161
9.1.1. General 161
9.1.2. Miscibility of polymers 161
9.1.3. Detection of miscibility 164
9.1.4. Block copolymers 165
9.1.5. The formation of dispersions 168
9.1.6. Properties of blends 171
9.2. Reinforcement by particles 176
9.3. Short fibres 177
9.4. Long fibres 181


10.1. Thermoplastics 185
10.2. Thermosets 189
10.3. Elastomers 190


11.1. Principles of processing 193
11.1.1. General 193
11.1.2. Production cost 195
11.1.3. Compounding 197
11.2. Casting and compression moulding 198
11.2.1. Casting 198
11.2.2. Rotational moulding 199
10 From polymers to plastics

11.2.3. Compression moulding 201

11.3. Injection moulding 205
11.3.1. General 205
11.3.2. Limitations 208
11.3.3. Defects 209
11.3.4. Foam 213
11.4. Calendering and extrusion 213
11.4.1. Calendering 213
11.4.2. Extrusion 215
11.4.3. Film blowing 219
11.4.3. Bottle blowing 220
11.5. Secondary shaping 222
11.5.1. Thermoforming 222
11.5.2. Cold sheet forming 224
11.5.3. Forging 224
11.5.4. Bending 225
11.5.5. Machining 226
11.5.6. Welding 226
11.5.7. Gluing 228
11.5.8. Surface coating 229
11.6. Processing of composites 230
11.6.1. General 230
11.6.2. Impregnating 230
11.6.3. Foaming 232




1.1. Origin of plastics / polymers

The main characteristic of polymers is, that they are composed of extremely large
molecules; their molecular mass ranges from 10.000 to more than 1.000.000 g/mol,
in contradiction to “normal” low-molecular substances, which in general are in the
order of 100 g/mol (water 18, sugar 342). Polymer molecules are often long, thread-
like chains, which are sometimes branched, sometimes chemically cross-linked with
each other so that they form a network.
Polymers are abundantly present in nature, in vegetable and animal tissues (mainly
as cellulose and proteins). Several technically used polymers have a natural origin;
they are being used as technical materials as they are harvested from natural
materials (“natural polymers”).
Other polymers are partly from a natural origin; the chain molecule has grown in a
living tissue, but has been chemically modified into a “half-synthetic polymer”.
A growing number of polymers is wholly synthetic; the chain molecule or the
network is being built up from small molecules (monomers) in a chemical process.
Some examples of these categories are given below:

Natural polymers:
– vegetable: timber, cotton, jute, sisal, hemp, cork, etc.
– animal : wool, silk, fur, etc.

Half-synthetic polymers:
– from wood: celluloid, cellophane, viscose-rayon, cellulose plastics,
– from milk: casein, from which casein plastics,
– from hides, via a tanning process: leather,
– from rubber latex, via i.a. vulcanization: technical rubber.

Synthetic polymers:
These are synthesized from low-molecular components, mostly organic monomers.
Most of the monomers are prepared from fossil fuels; here we can distinguish
between two main categories:
– carbochemistry (from coal)
– petrochemistry (from petroleum or natural gas).
12 From polymers to plastics

– Coal can be pyrolysed above 800 °C into coke, tar and a series of hydrocarbons.
– Gasification of coal with steam and air results in, a.o., a series of hydrocarbons.

From distillation of crude oil a number of fuels are obtained (kerosine, petrol,
gasoline, etc.) and a residue. The latter can be transformed into a series of lighter
components by vacuum distillation and thermal cracking. Also lighter fuels
(“naphtha”) can be converted to a series of various hydrocarbons by thermal cracking
and fractionation. Natural gas can supply a range of other components by steam
conversion or partial oxidation. It is important that thermal cracking of saturated
hydrocarbons results in unsaturated ones, containing one of more double bonds
between the C-atoms. This renders these molecules very suitable to be used as
monomers in polymerisation reactions. A long saturated chain is, for instance, split
up into a number of shorter molecules which are unsaturated due to a shortage of H-
atoms; they contain, therefore, one ore more double bonds:



Some of these molecules, such as the first and the second (ethylene and propylene)
can be used to build saturated chains (PE and PP). Other ones, such as the third one
(butadiene) form unsaturated chains, which can react with sulphur to form
vulcanized rubbers. (See Qu. 1.3, 1.4 and 1.5).

Polymer synthesis
From the monomers obtained from carbochemical or petrochemical sources
polymers can be built up. As mentioned above, double bonds can thereby play an
important role. Some simple examples:
C = C ethylene → –– C – C – C – C – – polyethylene, PE

C =C –C =C butadiene → –– C –C =C –C ––– polybutadiene

Polymers can also be synthesized from saturated monomers, e.g. by condensation of

an organic acid with an alcohol into an ester with the formation of water. Simple
alcoholes and fatty acids, such as ethyl alcohol (C2 H5 OH) and acetic acid
(C2 H5 COOH) result in a low-molecular ester, ethyl acetate, but when bifunctional
Introduction 13

molecules are used, a polymer chain (a polyester) is formed:

HOOC–R1 –COOH + HO–R2 –OH → ––CO–R1 –CO–O–R2–O–– + H2 O

When the alcohol is trifunctional (a triol), the result is a network:.

Figure 1.1.

Plastics as technical materials are based on polymers (or macromolecular
substances), but in most cases they contain a number of added components. Such an
added material may be another polymer; in this case we have a polymer blend.
Moreover, there is a large variety of additives and fillers, compounded into the
polymer for various purposes, which are roughly categorized below:
– in behalf of processing:
– lubricants for the transportation in the processing machine,
– antioxydants for protection of the polymer against oxydative degradation
during processing,
– sulphur, for the vulcanization of rubbers,
– accelerators, for speeding up the network forming reaction with rubbers and
– blowing agents, for producing foams,
– etc.
– in behalf of mechanical properties:
– plasticizers, e.g. in PVC, to obtain flexibility,
– quartz powder, mica, talcum etc. to improve the stiffness.
– short glass fibres, to improve stiffness and strength,
– rubber particles, to improve impact strength,
– etc.
– in behalf of other properties:
– ultraviolet stabilizers, to protect the polymer against degradation in sunlight,
– antioxydants, for protection against degradation during use at elevated
– antistatic agents, to reduce electrostatic charging,
– pigments,
– cheap fillers such as wood flour, for price reduction,
– flame retarding additives, etc
14 From polymers to plastics

1.2. Main categories

Taking the chain structure as a criterion, polymers can be subdivided into two main
categories, viz. single chains and networks.
Single chains are linear macromolecules, though the chains may be branched. An
average scale model for such a polymer chain is a human hair with a length of a
metre. Normally, chains are not present in this extended shape, but rather as diluted
coils, which may, in the scale model, have a diameter of a few cm.
The coils are mutually entangled; this fact is largely responsible for the special
behaviour of polymers! Between chain segments relatively weak interaction forces
are present; chain segments can move with respect to each other under the influence
of relatively small external stresses. Consequently, the stiffness of polymers is rather
Flow, on the contrary, is strongly hindered by the entanglements between the coils;
this is why fluid polymers show excessively high viscosities, which necessitate the
use of heavy processing machines.
In networks the molecular chains are connected by strong primary chemical links; in
fact a network is a single gigantic molecule. Networks can be formed in two different
a. by forming bridges between single chains; this is the case with the vulcanization
of rubbers, mostly by sulphur bridges. But also curing of unsaturated polyester
resins is a matter of forming bridges between chains, in this case with the aid of
b. by reacting bivalent with tri-(or more)-valent molecules, e.g. in synthesizing
formaldehyde resins.
In networks chain entanglements and local movements of chain segments with
respect to each other also occur, but flow is not possible.
From the foregoing we can divide the field of macromolecular materials into three
main categories:
Thermoplastics are non-crosslinked systems, which flow at elevated temperatures
and which, upon cooling, return to the solid state.
Synthetic elastomers are analogous to thermoplastics, but they are in a softened
condition. As such they show flow, but after the formation of a network (by cross-
linking) they are no longer fluid but retain their shape.
Thermosets owe their name to the fact that the formation of a network, the curing
reaction, in many (but not in all) cases occurs at elevated temperature. The network
is considerably tighter than in vulcanized rubbers. A formed thermoset does not
increase in hardness or stiffness upon temperature increase, but, on the contrary,
shows softening, but no flow.
Introduction 15

Within each of these three categories a large number of main types exists, each
characterized by a a specific structure of the macromolecule. Within each main type
several variations occur, e.g. concerning chain length, chain regularity, copoly-
merisation (presence of more than one monomer in the chain), etc.
In the assortment of technically used materials we also meet a broad variability of
additives and fillers, added by the manufacturer.
In the next section an (incomplete) survey is given of a number of polymers, as a
first orientation in the field of plastics materials.

1.3. The most important plastics

(See Qu. 1.14 – 1.20).

1.3.1. Thermoplastics
Polyethylene (PE) is a rather soft and tough, crystalline polymer, which is being
manufactured in three main types: LDPE (low-density PE ≈ 0.92 g/cm 3 ), HDPE
(high-density PE ≈ 0.95 g/cm3 ), and, more recently, LLDPE (linear-low-density PE,
≈ 0.92 – 0.95 g/cm3). The stiffness of PE increases strongly with increasing density.
All types gradually lose their properties upon temperature increase, and they melt at
105 to 130 °C, respectively. Principal applications are: packaging film, bags, pipes,
crates, pails, bottles, etc. Special grades are made in smaller quantities, such as
UHMPE (ultra-high-molecular), which is extremely tough and abrasion resistant, and
which is, moreover, used for making super-strong PE-fibres. A new development is a
series of PE with very low density (.86 to .90 g/cm 3 ), a series which extends into the
region of rubbers.
Polypropylene (PP) resembles PE, but is somewhat harder and stiffer than HDPE. It
is crystalline as well, and melts at ≈ 165 °C. Its impact strength at lower temperatures
is quite poor; therefore PP is often modified with a certain amount of rubber (mostly
built-in as a copolymer). Main applications are: film for packaging, fibres, crates,
pipes, automotive parts (often with reinforcing fillers). A special feature of PP is its
ability to form integral hinges with a practically unlimited resistance against repeated
Polyvinylchloride (PVC) is a hard, amorphous polymer which softens at about 85 °C.
Also in PVC rubbers are sometimes added in order to improve the impact strength.
The main applications of PVC are: pipes, gutters, front panels of buildings, cables,
bottles, floor tiles. A much softer and more flexible material is obtained by blending
with plasticizers: soft or plasticized PVC is being used in artificial leather, tubes and
hoses, footwear, films, etc.
Polystyrene (PS) is an amorphous, very brittle, hard polymer with a softening
temperature of about 90 °C. Improvement of its impact strength is again obtained by
16 From polymers to plastics

blending with rubber (in most cases butadiene rubber), which goes at the cost of
stiffness. Unmodified PS is being largely applied as foam for packaging and thermal
insulation. High-impact PS (TPS or HIPS) is used for coffee cups, household ware
Styrene-Acrylonitrile (SAN) is somewhat stiffer than TPS, and has a better resistance
against impact and temperature. It is mainly used in parts of household- and
electrical appliances, battery houses etc.
Acrylonitrile-Butadiene-Styrene (ABS) is sometimes a terpolymer of three mono-
mers, but in most cases a blend of two copolymers. ABS has an excellent impact
strength and a relatively high softening temperature (about 110 °C). Its stiffness is
only marginally lower than that of PS. It finds large-scale applications in the
automotive industry, in toys, telephones, TV-housings, etc.
Polymethylmethacrylate (PMMA), carrying trade names such as Perspex and
Plexiglass, is an amorphous, relatively hard and transparent polymer. Its stiffness is
retained until near its softening temperature (110 °C). Most applications are based on
its superior optical qualities: safety glass, decoration material, traffic signs, etc.
Polyamide (PA) is generally known as “nylon”. This is a collection of polymers,
which differ in chain structure, and which are, according to the numbers of
consecutive C-atoms in the chain, designated as (a.o.) PA-6, PA-6,6, PA-11, PA-4,6
and PA-12. Initially PA’s were only used as fibre materials, but later on they found a
position under the engineering plastics. Polyamides are crystalline polymers with
relatively high melting points (between about 200 and 300 °C). They possess a good
impact strength, due to the fact that they absorb several percents of water from the
atmosphere. Moreover, a good abrasion resistance and a low friction make them
suitable for technical use, such as in bearings and gear wheels. Quite often
polyamides are reinforced with short glass fibres to improve their stiffness.
Polyoxymethylene (POM) is, again, a crystalline polymer, with a melting point of
about 180 °C. Its mechanical properties enable it to gradually replace metals in a
number of applications. Many technical parts are being made from POM, such as
gear wheels, bars, automotive accessories, parts of several apparatuses and machines.
The polymer is used as such (e.g. “Delrin”), but also as a copolymer with a small
amount of ethylene oxide (e.g. “Celcon” and “Hostaform”).
Polycarbonate (PC) is, up to about 140 °C, an amorphous glassy transparent polymer
with excellent mechanical properties, in particular as regards its impact strength.
This renders it very suitable for substitution of glass, but also for a series of technical
applications in which it replaces metals. For the latter, reinforcement with short glass
fibres opens further possibilities. A weak point of PC is its poor resistance against
environmental stress cracking in contact with a number of organic liquids.
Introduction 17

Polyethylene terephtalate (PETP) is a (saturated) polyester and is, like nylon, known
for its large-scale use as a textile fibre. Moreover it is being applied at an increasing
scale as a plastic, viz. in films, bottles (the “PET-bottle”) and injection-mouldings.
Though its stiffness decreases significantly above 70 °C, it remains a solid up to its
melting point (255 °C).
Polybutylene terephtalate (PBTP) differs in its chemical structure only slightly from
PETP; its melting point is somewhat lower and its processability is better. The
applications in injection moulded articles are similar to those for PETP.
Polyphenylene oxide (PPO) or Polyphenylene ether (PPE) is an amorphous polymer
with a softening temperature of about 210 °C. To improve its processability it is
mostly blended with PS (modified PPE, e.g.”Noryl”), which is at the cost of its heat
distortion temperature. The properties are excellent; the applications are mainly in
fine-mechanical construction, in automotive parts, in household equipment etc.
Polysulphone (PSU) is a high-performance polymer with superior mechanical,
electrical and thermal properties in a temperature region from –100 up to 180 °C. It
is mainly used in exacting mechanical and electrical applications.
Polyphenylene sulphide (PPS) (e.g.”Ryton”) is a highly crystalline polymer with a
melting point of 290 °C. It combines good mechanical properties with very high
thermal and chemical resistance; it is, moreover, self-extinguishing. It is, i.a., used as
protective coating on metal surfaces.
Polyimide (PI) caps all other polymers in its temperature range of use (–200 to 260
°C in air; short-time even up to 500 °C). Because of its high price, it is used in
special cases only, such as space vehicles, nuclear reactors and some electronic parts.
Newer developments, related to polyimide, are the polyether imides (e.g. “Ultem”),
polyester imides and polyamide imides (e.g. “Torlon”), all with very good
mechanical, thermal and electrical properties and self-extinguishing.
Polytetrafluoroethylene (PTFE), mostly known as “Teflon”, has as special properties
its high melting point (327 °C), its very good resistance against chemicals and its
extremely low friction. though the polymer is mechanically weak and shows a strong
tendency to creep, though processing is very difficult (only possible via a sintering
process) and though it is very expensive, it is being used in a number of applications
such as bearings, pipes, sealing rings, electrical insulation and coatings on kitchen
pans. Often the polymer is reinforced with fillers.
Tetrafluoroethylene-perfluoropropylene (FEP) resembles PTFE in its properties, but
can be processed as a thermoplastic polymer.
Polyvinylidene fluoride (PVDF) and Ethylene tetrafluoroethylene copolymer (ETFE)
can be considered as “diluted” PTFE’s, which in their structure and their properties
18 From polymers to plastics

are in between PTFE and the polyolefins PE and PP. They are processable with the
normal techniques and find similar applications as PTFE.
Cellulose acetate (CA) and Cellulose acetate butyrate (CAB) are, contrary to the
polymers mentioned so far, not fully synthetic, but derivates of vegetable cellulose.
They are strong, tough and well processable materials, used in many household- and
technical applications, such as hammer heads, magnetic tape, toys etc. CAB has a
higher form stability than CA, and is used in automotive accessories and in pipes.
Polybutylene (PB) belongs to the family of polyolefins (PE and PP), but is much less
used because of its higher price. Though it does not differ very much from PE in
stiffness and melting point, it has a better resistance against creep and environmental
stress cracking than PE and PP and a higher toughness and tear resistance. Its
applications are mainly in film for heavy-duty bags and in hot water transport pipes.
Polymethylene pentylene (PMP) is, again, a polyolefin, but with a much higher
melting point than the other ones (240 °C). Despite its crystalline nature PMP can be
well transparent. It is used a.o. in laboratory bottles. One of its trade names is
Polyether ether ketone (PEEK) and Polyether sulphone (PES) belong to the most
recent developments in the field of technical high-performance polymers. Both
possess very good thermal and mechanical properties, which can be further improved
by reinforcing fibres. Their application is mainly in aircraft and space vehicles.
Polyketone (PK) (“CARILON”) is a new polymer with an attractive combination of
properties (very tough, high abrasion resistance, high temperature resistance, easy to
process, good chemical resistance and barrier properties). It is trying to find its way
in various potential applications.

1.3.2. Thermosets
Phenol-formaldehyde (PF) was the first fully synthetic macromolecular material
(“Bakelite”, 1907). In a slightly precured condition and provided with fillers, it is, as
a moulding powder, available for processing into end-use articles such as bulb
fittings, switch housings, coils, laminated wood and foam for thermal insulation.
Ureum-formaldehyde (UF) is comparable to PF, but is somewhat stiffer and has,
because of its colourlessness and gloss, a more attractive appearance. The
applications are in the same fields as PF.
Melamine-formaldehyde (MF) is qualitatively better than UF. It is, therefore, used in
more demanding applications such as crockery, various electrotechnical articles and
decorative panels.
Unsaturated polyesters (UP) can, with a second component, e.g. styrene, and with
Introduction 19

initiators and accelerators be cured into a network structure. The reaction can take
place at room temperature. UP is, in most cases, used in combination with glass
fibres, and finds its applications in pipes, vessels, boat building etc.
Epoxy resin (EP) has to be mixed with a second component, the curing agent, to
undergo the curing reaction, which, as with UP, can take place at ambient
temperatures. Epoxy/fibre composites (with glass, carbon or aramid fibres) find
similar applications as UP/glass, but are being applied more selectively because of
their higher price and better properties. Besides, epoxies are used in lacquers and
adhesives and as casting resins in electrotechnical applications.
Polyurethanes (PU). The thermosetting type of this large family of polymers is
mainly used as foam. A mixture of two components with a foaming agent forms a
light, hard foam, which is a superior thermal insulator.

1.3.3. Synthetic elastomers

Styrene butadiene rubber (SBR) is, quantitatively, the most important synthetic
rubber. It is a copolymer of styrene and butadiene in such a ratio that its rubbery
nature predominates. vulcanization is carried out with sulphur, reinforcement with
carbon black. It is used at a very large scale in tyres for passenger cars, thanks to its
excellent combination of abrasion resistance and friction on the road. In large tyres it
can not replace natural rubber because of its heat development (hysteresis losses).
Butadiene rubber (BR) or Polybutadiene has an excellent abrasion resistance and a
very low damping, but is, undiluted, too “jumpy” for use in tyres. In blends with
SBR or natural rubber a good compromise of properties can be obtained.
Isoprene rubber (IR) or Polyisoprene is a synthetic copy of natural rubber (NR) and
approaches NR in its properties. Besides for tyres, IR is, because of its good flow
properties, suitable for injection moulding.
Butyl rubber (IIR) is derived from polyisobutylene, a polymer which is not further
mentioned in this chapter, which has a rubbery nature, but which can not be
vulcanised in the conventional way with sulphur. This objection is taken away by
copolymerisation with a small amount of isoprene. Butyl rubber has a very low
resilience, but outrivals all other rubbers in resistance to gas permeation; for that
reason it is generally used for tyre inner tubes.
Chloroprene rubber (CR) is a synthetic rubber with very high chemical resistance,
and is, therefore, applied in cable protection, oil transport tubes etc.
Nitrile rubber (NBR), a copolymer of butadiene and acrylonitrile, is characterized by
its high resistance against light and oxygen, thus against ageing. Moreover it is not
attacked by oil and several organic solvents. For these reasons, it finds its place in
demanding applications.
20 From polymers to plastics

Ethylene-propylene rubber (EPR or EPDM) is, basically, a copolymer of ethylene

and propylene. Because of the random arrangement of the monomers in the chain,
crystallization does not occur, and the material behaves as a rubber. Just as with
polyisobutylene, vulcanization with sulphur is impossible (the chain is saturated).
Also here, a small amount of another monomer is incorporated, which enables the
vulcanization and thus the use as a technical elastomer. EPR has a high resistance
against ageing and chemical attack, and is, compared with other “specialty” rubbers,
relatively cheap.
Silicone rubbers have, contrary to all other polymers, no carbon atoms in their main
chain, but silicium and oxygen atoms only. They can be used up to very high
temperatures (250 °C) and are highly resistant against ageing, so that, despite of their
high price, they are frequently used in demanding applications.
Thermoplastic elastomers (TPE’s) are characterized by the exceptional property that,
without vulcanization, they behave as cross-linked rubbers. They are block-
copolymers, in which blocks of the same nature assemble in hard domains, acting as
cross-links between the rubbery parts of the chain. These hard domains lose their
function when they reach their softening temperature, so that the material can then be
processed as a thermoplast. One of the oldest member of the family of TPE’s is SBS
(styrene-butadiene-styrene block copolymer), but several other TPE’s have been
developed, i.a. on the basis of polyesters, polyurethanes and polyolefins. In their
properties these polymers cover a broad range between conventional rubbers and soft
Polyurethane rubber (PUR). Not only in the thermosets (and the thermoplastics), but
also in the field of synthetic elastomers polyurethanes have found a position, namely
as a softer type. It is, again, formed from two components and is, with a blowing
agent, processed into a foam. Polyether mattresses belong to this category, but also
microcellular structural foams, used in bumpers, head- and arm-rests in motorcars,

1.3.4. Composite plastics

Blends of polymers are manufactured and applied at an increasing scale. Only in
exceptional cases are polymers soluble in each other and can form a homogeneous
blend (an example: PPE + PS, a blend known as “Noryl”). In most cases blends are,
therefore, dispersions. Rubber particles are dispersed in brittle polymers to improve
their impact strength (toughened PS and PP, ABS etc.), but also hard polymers are
combined to reach a favourable compromise between properties (and price).
Reinforcement with particles such as chalk, quartz, mica and glass spheres, is
frequently carried out with thermoplastics and thermosets to obtain a higher stiffness
(and sometimes a higher strength). There is a gradual transition from high-quality to
Introduction 21

cheap fillers, the latter being mainly used as price reducing agents, but also for. e.g.
reducing shrinkage in processing. Rubber vulcanisates gain considerably in strength
and abrasion resistance by the incorporation of carbon black (up to 40 weight %)
Reinforcement with short fibres is important for thermoplasts and thermosets. With
the former, very short fibres (glass or other) are blended into the polymer; the latter
allow the use of longer fibres. The effect is a 3- to 5-fold increase of the stiffness and
a 1.5 to 3-fold increase in strength.
Foams can be made from thermoplastics, thermosets and rubbers. Densities can be
obtained from nearly solid down to 200 times diluted. Structural (or integral) foams
have a solid skin. The best known foam materials are polystyrene foam,
polyurethane foam and polyether foam.
Reinforcement with continuous fibres. In this case the fibres, as strands or cloth, are
largely responsible for the mechanical properties. Traditionally, thermosetting resins
are used in this case, because of the ease of impregnating fibre bundles or cloth with
the low-viscosity uncured resin. Recently however, techniques have been developed
to also reinforce thermoplastics with long fibres. Conventional combinations are
polyester/glass (GF-UP) and epoxy/glass (GF-EP), but also high-quality fibres as
carbon and aramide, and thermoplastic matrices such as PEEK are being used for
special applications.

Molecular composition

2.1. Chain structure

A linear chain consists of a “backbone”, the main chain, at which side groups are
attached. In this section a simple classification of the various types of main chain
will be given, followed by a survey of frequently occurring side groups.

2.1.1. Main chain (Qu. 2.1 and 2.2)

Carbon atoms only:
– saturated:
example: polyethylene (PE)
––[– C – C –] ––
H H n
also PP, PS, PVC, PB, PMMA, PTFE, etc. (see § 2.1.2)
– unsaturated:
––C–C=C– C–C–C= C–C––
example: polybutadiene
––[– C – C = C – C –]n ––
also IR, CR, etc. (see § 2.1.2)
– more unsaturated:

––C= C–C= C–C= C––

example: polyacetylene.

––[– C = C –]n ––
Carbon and oxygen atoms:
––C–O– C–O– C–O– C–O––
Molecular composition 23

example: polyoxymethylene (POM) :

––[– C –O–] ––
––C–C–O– C–C–O– C–C–O––

example polyethyleneoxyde (PEO)

––[– C – C –O] ––

Carbon and nitrogen atoms:

––C–C–C–N– C–C––

example: various types of PA (nylon):

|| || | |
––[–C –(CH2 )p –C –N–(CH2 )q –N–]n ––

Carbon, oxygen and nitrogen atoms:

––C–C–––––N– C–O– C–C––

example: various types of polyurethanes:

|| | | ||
––[–(CH2 )p –O– C –N–(CH2 )q –N–C –O–]n ––

Carbon rings:

–– ––

example: polycarbonate (PC):

| O
––[–O– – C– –O– C –]n ––

other example: polyethylene terephtalate (PETP):

|| ||
––[–C – –C –O– CH2 –CH2 –O–]n ––
24 From polymers to plastics

also PBTP, with four instead of two CH2 groups,

|| ||
––[–C – –C –O– (CH2 )4 –O–]n ––

polyetherether ketone, PEEK:

––[–O– –O– –C – –]n ––

polyphenylene ether, PPE:


––[– –O–] ––

Combination with sulphur atoms:

polyphenylene sulphide (PPS):
––[– –S–]n ––

polysulphone (PSU
| ||
––[– – C– –O– – S– –O–]n ––
| ||

polyether sulphone, PES:

––[–O– – S– –]n ––

Silicium and oxygen atoms only:

example: polydimethylsiloxane (silicone rubber):
––[–Si –O–] ––
| n
Molecular composition 25

Multiple rings:
example: polyimide (PI):
––[–N N– –O– –] ––

2.1.2. Side groups

– The most frequently occurring side group is the hydrogen atom; it is the only one
with polyethylene, polybutadiene and polyoxymethylene.
– Other arrangements often met are:

– vinyl polymers
––[– C – C –]––
with: R = CH3 polypropylpene (PP)
R = Cl polyvinylchloride (PVC)
R = C6 H 5 = – polystyrene (PS)
R = CH2–CH3 polybutylene (PB)
R = CN polyacrylonitrile (PAN)

– vinylidene polymers
––[– C – C–]––
with: R = Cl polyvinylidenechloride (PVDC)
R=F polyvinylidenefluoride (PVDF)
R = CH3 polyisobutene (PIB)

– polytetrafluorethylene (PTFE)
––[– C– C–]––
with: R=F

– polymethylmethacrylate (PMMA)
H R1
––[– C – C –]––
H R2
with: R1 = CH3 and R2 = COOCH3
26 From polymers to plastics

– polydienes
––[– C – C = C– C –]––
with: (R = H polybutadiene (BR)
R = CH3 polyisoprene (IR)
R = Cl polychloroprene (CR)
– etc. etc.

2.1.3. Copolymers
Copolymer chains are built up from more than one type of monomeric units. Some

ethylene + propylene: EPR

styrene + butadiene: SBR or SBS (see § 2.3)
styrene + acrylonitrile: SAN
isobutylene + isoprene: IIR (butyl rubber)

Also terpolymers (with three monomers) are made:

ethylene + propylene + a diene: EPDM

acrylonitrile + styrene + acrylic ester: ASA
ethylene + CO + propylene: PK, poly ketone (“Carilon”).
(see further § 2.3)

2.2. Chain length and distribution

2.2.1. Averages
The chain length can be expressed as degree of polymerisation P (number of
monomeric units in the chain), or, in most cases, as molar mass (or molecular
weight) in g/mol. Sometimes the “better” unit, kg/mol is used; the difference is a
factor of 1000. In this book the older unit, g/mol, will mostly be used.

An example: PE with a degree of polymerisation P of 5000 has a molar mass:

M = 5,000·(2C + 4H) = 5,000·(24 + 4) = 140,000

Of most polymers various types exist, which differ in molecular mass. The strongest
example is PE, which is made at more than 10 different levels of M, the extremes
differing by a factor of 20. But also within each type the chains may largely differ in
length, namely by a factor of 100 or even 1000. Therefore every statement about a
molar mass indicates an average.
Molecular composition 27

Averages can be defined in several ways, for example after number or after weight.
A simple example:

1 chain with mass 100 and

1 chain with mass 10

then the number average is:

1 1
· 100 + 2 · 10 = 55
since the numbers are equal, and the number fractions n1 and n2 are both 2.
Now consider a system with:

1 chain with mass 100 (total mass 100) and

10 chains with mass 10 (total mass 100)

then the weight average is:

1 1
· 100 + 2 · 10 = 55
because now the weight fractions are equal (w1 en w2 both 2 ).

Expressed in a formula:

number average Mn = ∑ ni·Mi

in which ni is the number fraction of the chains with mass Mi , and ∑ ni = 1.

The weight fraction is Wi = ni·Mi ; now however, ∑ Wi is not equal to 1; the fractions
Wi must, consequently, be reduced (normalized) to wi, so that ∑ wi = 1:
ni ·Mi Wi
wi = ∑ n ·M = ∑ W
i i i

The weight average is now:

Mw = ∑ wi·Mi

Also valid is:

— ∑ ni ·Mi 2 ∑ ni ·Mi 2
Mw = ∑ n ·M = —
i i M n

In most cases it is necessary to work backward from the weight fractions wi:
Ni = M
28 From polymers to plastics

∑ Ni ≠ 1

so Ni has to be normalized to ni with:

Ni wi /Mi
ni = ∑ N = ∑ (w /M )
i i i
— —
Now ∑ ni = 1 , and Mn follows from Mn = ∑ ni·Mi .

The calculation of M n is easier with:
— ∑ wi 1
Mn = ∑ ni·Mi = =
∑ (wi /Mi ) ∑ (wi /Mi )

Also “higher” averages are used:

Mz = ∑ zi ·Mi
in which
wi ·Mi
zi =
∑ wi ·Mi (normalized)
— ∑ wi ·Mi 2 ∑ ni ·Mi 3
Mz = ∑ w ·M =
i i ∑ ni ·Mi 2

This is the so-called z-average. A further step in the same direction gives the (z+1)-
— ∑ ni ·Mi 4
Mz+1 =
∑ ni ·Mi 3
— — —
We apply the formulae for Mn . Mw and Mz to the simple example given above:

A 100 10 (100 + 10)

B 100 10 (100 + 10 × 10)
M1 = mass chain 1 100 100
M2 = mass chain2 10 10
n1 = number fraction 1 1/2 1/11
n2 = number fraction 2 1/2 10/11

Mn = ∑ ni·Mi 55 18.2
w1 = weight fraction 1 = n1·M1 /∑ ni ·Mi 10/11 1/2
w2 = weight fraction 2 = n2·M2 /∑ ni ·Mi 1/11 1/2

Mw = ∑ wi·Mi 91.8 55
z1 = z-fraction 1 = w1 ·M1 /∑ wi ·Mi 100/101 10/11
z2 = z-fraction 2 = w2·M2 /∑ wi ·Mi 1/101 1/11
Molecular composition 29

Mz = ∑ zi ·Mi 99.1 91.8

Another, more realistic example: A blend is made of three monodisperse fractions

(i.e. in each fraction all chains have the same length), with masses of 20, 50 and 30
grammes and molar masses of 20,000, 100,000 and 300,000 g/mol, respectively. The
mass fractions are now:
w1 = 100 = 0.2

w2 = 100 = 0.5

w3 = 100 = 0.3 (∑ wi = 1)

This leads to:

Mw = 0.2·20000 + 0.5·100000 + 0.3·300000 = 144000

Mn can be calculated with:
0.2 0.5 0.3 4.8
N1 = 20 000 , N2 = 100 000 , N3 = 300 000 , ∑ Ni = 300 000

After normalization we find: n1 = 10/16, n2 = 5/16, n3 = 1/16.

Mn = (10/16)·20000 + (5/16)·100000 + (1/16)·300000 = 62.500

(The calculation is simpler with Mn = 1/∑(wi /Mi )).

The z-fractions are, unnormalized:

Z1 = 0.2·20000 Z2 = 0.5·100000 Z3 = 0.3·300000

and normalized:
4 50 90
z1 = 144 z2 = 144 z3 = 144

It simply follows now that:

Mz = 222,778
— — —
The examples show clearly that in Mn , Mw and Mz in this order of sequence, the
longer chains play an increasingly dominant part.
In most cases M w is used for general characterization of the molar mass besides D =
— —
Mw/Mn , the heterogeneity index, is of importance as a simple measure for the spread
in chain length. In the first example D = 1.67 for A and 3.03 for B. In the second
example D = 2.3. For technical polymers D can have values from 1.1 up to 30.
30 From polymers to plastics

An even more realistic question is, how the averages change when two polydisperse
polymers of the same type are being blended. Suppose two batches, A and B, are
— —
available with weight average molar masses (Mw)A and (Mw)B and a blend is made
of a weight fraction ϕ of A with a weight fraction (1 – ϕ) of B. For the blend C the
weight average is now:
— — —
(Mw)C = ϕ·(Mw)A + (1 – ϕ)·( Mw)B.

The number of chains in 1 gramme of polymer is N/Mn (N is the Avogadro number).
In 1 gramme of the blend C, ϕ grammes of A and 1 – ϕ grammes of B are present,
— — —
N/(Mn )C = ϕ·N/(Mn )A + (1 – ϕ)·N/(Mn )B,
— — —
ϕ/( Mn )C = 1/(Mn )A + (1 – ϕ)/(Mn )B

How broad is now the distribution of the blend C? We find D C by multiplying the
— —
expressions for (Mw)C and 1/(Mn )C. A much simpler expression is found when we
consider a special case, viz. two components with the same degree of dispersion D,
— — — —
in which (Mw)A = α·(Mw)B (and, therefore, (Mn)A = α·(Mn)B). Some calculation leads

(Mw)C (α – 1)2
DC = = D[ϕ ·(1 – ϕ) · + 1].

(Mn)C α
— —
An example: α = 4, Mw of A (e.g. 60,000) is four times greater than Mw of B
1 —
(15,000). We choose ϕ = 2 , so we blend equal masses, and the blend has an Mw of

37.500. When now, for instance, D = 6, so that the Mn’s of both components are six

times smaller than the Mw’s (10,000 en 2,500, respectively), then
1 1 9 25
DC = 6·[ 2 · 2 · 4 + 1] = 6· 16 = 9.375 ;

the distribution width is increased by more than a factor 1.5. Mn of the blend is now
37,500 / 9.375 = 4,000.
This simple rule, though based on the assumption that the molar mass distributions
are similar, has practical significance because polymers made in one and the same
process, do in general not differ much in their heterogeneity index.

So far, we have considered the magnitudes of several types of average molar masses.
The question may arise what the use is of these different averages; the attention paid
to them might look a bit artificial. Well, it is easy to understand that various polymer
properties are influenced by the molar mass. Moreover, it appears that different
physical properties of polymers are controlled by different types of average molar
mass. A few examples:
Molecular composition 31

– The number average, Mn , is proportional to the mass of one gramme of the
polymer divided by the number of chains. It is, therefore, a function of the
number of chains, and thus of the number of chain ends (apart from branching).
Now suppose that a crack with a sharp tip propagates through a piece of the
polymer, then this propagation is easier as less chains have to be broken, and it
meets more chain ends on its way. This means that all properties in which crack
propagation plays a role, such as impact strength, tear strength and environmental

stress cracking are governed by Mn .

- The weight average, Mw , is the most practical one. If you blend, for instance,
equal quantities (in kg) of two batches with molar masse M1 and M 2 , then you
would expect the average (M1 + M2 ) / 2. And this is, of course, a weight average,
for nobody would consider to blend equal nunbers of chains with each others! In

addition, Mw is specially responsible for the viscosity of the polymer in molten
condition, and, therefore, for the processability. This holds, however, only as a
first approximation, as we will see later in § 5.3.

– As a third example we look at the z-average, Mz. In the foregoing we have seen
— —
that Mz (and, in particular, Mz+1 ), are governed by the longest cahains, while the
sort chains have hardly any effect. A property in which the longest chains have a
predominant effect, is the melt elasticity. This is a rather unexpected expression,
because a fluid is, in general, not an elastic substance. Yes, but for polymers, it is!
Let us consider a polymer, consisting for the larger part of short chains, with a
single very long chain. When this polymer is deformed in a flow field, the long
chain will, as a tracking-thread, keep the short chains together, and force them to
return to the most probable conformation, so to return back. This is further treated
in § 5.3.2. This “melt-elasticity” is of major importance for practical processing
technology and for some other properties. It is clear that we cannot neglect higher
— —
averages, such as Mz and Mz+1

2.2.2. Example of a chain length distribution

In a polymerization process the chain length distribution or molar mass distribution
(MMD) is influenced by a large number of factors and conditions; the kinetics of the
reaction plays a very important role. The calculation of the resulting MMD is thus
very complicated. For one of the simplest cases, a step reaction with
polycondensation, a first-order approach is given here. As an example we take a
hydroxy acid HO–R–COOH, which, upon condensation, forms the chain –[– O–R–
CO–]n .

With each step a –COOH and an –OH group react with each other, forming an ester
group and a water molecule. If we denote the numbers of –COOH and –OH groups
both by U, then at the start of the reaction U = U0 = the total number of hydroxy acid
32 From polymers to plastics

After some time t, U = U(t); the number of –COOH and –OH groups disappeared is
(U0 – U), which also is the number of ester groups formed. We define the conversion
grade p as the fraction of the number of groups which have reacted:
U0 – U
p= U0 or U = (1 – p)U0

Now there are U molecules, containing together U0 basic units; the number of units

per chain, the average degree of polymerization P, is, therefore:

— U0 1
P = (1 – p)·U = 1 – p
— —
An example: With a conversion grade of 0,99 P = 100. Evidently, this P is a number

average degree of polymerization, P n .

The question is now, how the distribution of P (and thus of M) looks like. We
consider, during the reaction, when the conversion grade is p, a single –COOH
group. The chance that the –OH group of this monomer has reacted with another
–COOH group, is p; the chance that it has not reacted, is (1 – p).

The chance of HOOC–R–OH thus is (1 – p)

The chance of HOOC–R–OOCR is p
The chance of HOOC–R–OOCR–OH is p(1 – p)
The chance of HOOC–R–OOC-R–OOCR is p2

The chance of (i – 1) times addition is pi – 1

the chance of not reacting with the ith monomer is (1 – p)
the chance of both events, so the chance of the occurrence of a chain of i units,
is the product pi – 1 (1 – p). This is, therefore the number fraction, ni.
Ni i – 1 (1 – p)
N = ni = p

It follows easily that:

∑ ni = 1

With this formula the value of the number average, found before, can be checked:
— _ ∑ ni ·i 1 – p 1 + 2p + 3p2 + … 1
Pn =i= ∑n =1–p· =1–p
i 1 + p + p2 + …

The weight average degree of polymerization is:

Molecular composition 33

— ∑ Wi ·i ∑ ni ·i2 ∑ pi – 1 (1 – p)i2 1 + 4p + 9p2 + 16p3 + …

Pw = = = =
∑ Wi ∑ ni ·i ∑ pi – 1 (1 – p)i 1 + 2p + 3p2 + 4p3 + …
(1 – p)3 1+p
= =
1 1–p
(1 – p)2
— —
The ratio D = P w/ P n , the heterogeneity index, is D = (1 + p). For the usually high
conversion rates (e.g. p = .995), D ≈ 2.

The chain length distribution can be represented graphically in several ways, e.g. as
number fractions or as weight fractions. In Figure 2.1 both cases are given (together
with the z-fractions) for a p-value of 0.99. From the graph of number fractions it
appears that the monomer has, numerically, the highest fraction! The distribution of
the weight fractions wi has its maximum at the number average degree of
— — —
polymerization, P n = 100, while P w = 199. Further calculation gives: P z = 299.

p = 0.99 0.004


100 200 300 400 500
Pn Pw Pz

Figure 2.1. Chain length distribution of a polycondensate.

2.2.3. Measuring methods

The number average Mn can, in principle, be determined by counting the molecules
in a gramme of polymer. This is possible by measuring colligative properties of the
polymer in solution; these are properties which are strictly dependent on the number
of molecules per unit volume of the solution, and independent of their nature or size.
Colligative properties are:
34 From polymers to plastics

- vapour pressure reduction,

- freezing point reduction,
- boiling point increase,
- osmotic pressure

For polymers with normal, high molar masses the osmotic pressure is the only
possibility; the other methods give a too small effect because of the relatively small
number of molecules.

The osmotic pressure, ∏, is the pressure difference between solution and pure
solvent when these are separated by a semipermeable membrane (see Figure 2.2).
For very dilute solutions the osmotic pressure is given by (see Qu. 2.35 and 2.36):
∏ = kT V

in which N is the number of molecules in a volume V, k is Boltzmann’s constant and

T the absolute temperature. N can, via the molar fraction and the weight fraction, be
expressed in the concentration c (g/dl) and the molar mass M; the formula then reads:
∏ = RT —

However, ∏/c does not automatically lead to the correct value of Mn , since ∏/c is
still slightly dependent on c. Measurements are, therefore, carried out at a number of
different concentrations and a plot of ∏/c against c is extrapolated to c = 0.

solution solvent

Figure 2.2. Principle of osmometry.

For the determination of the weight average, Mw , about light scattering is the
standard method. This is based on the fact that a density fluctuation in a solution,
brought away by the presence of a coil molecule, causes a deviation of a ray of light.
The amplitude of the scattered waves is proportional to the mass of the particle; the
intensity is proportional to the square of the amplitude and thus to the square of the
mass. From this simple reasoning it can be intuitively understood (though not
proven), that the total amount of scattered light from all molecules present is related
to the weight average molar mass.
A strict condition is, that the scattered waves do not interfere with each other so that
the concentration must be very low. Light scattering is a very laborious and difficult
method; minor contaminations are disastrous for the result. It is only used when
Molecular composition 35

absolute determinations of Mw are necessary.

Higher averages can be obtained by measuring the rate of sedimentation in solvents

with the aid of an ultracentrifuge. This is based on the fact that the rate of
— —
sedimentation depends on the molar mass. The measurements supply Mw and Mz,,

sometimes Mz+1 .
— — — —
The viscosity average, Mv , does not belong to the series Mn , Mw , Mz etc. The
measurement does not supply an absolute value of the molar mass, but can, by
contrast with the other methods mentioned so far, be carried out in an easy way. Its
principle is, that a polymer, even at very low concentrations, brings about a
significant viscosity increase of the solvent. This increase is not only dependent on
the concentration but also on the molar mass. The measurement is carried out as
follows (see also Qu. 2.14 - 2.18 and 2.31 - 2.34):

In a capillary viscometer the time is determined in which the liquid has dropped from
the initial level h1 to h2 (see Figure 2.3); this time is proportional to the viscosity of
the liquid. When we denote the viscosity of the solvent by η0 and that of the solution
by η, then the specific viscosity is defined as:

η – η0
ηsp =

This is the relative increase in viscosity, which, in a first approximation, appears to

be proportional to the concentration c. A better measure is, therefore :

ηsp η – η0
c = η0·c = ηred,

the reduced viscosity.

This ηred is still somewhat dependent on the concentration. It is, therefore, measured
at a number of (low) concentrations; the values found are extrapolated to c = 0,
where the intrinsic viscosity is read off, which is denoted by [η]:

h1 ηred



Figure 2.3. Principles of measuring intrinsic viscosity

36 From polymers to plastics

η – η0
[η] = lim ηred = lim
c→ 0 c→ 0 η0·c

[η] does not have the dimension of a viscosity, but of c–1; in most cases it is given in
A strongly schematized example: For a certain polymer in a certain solvent, times of
flow through the capillary at several concentrations are found as indicated in the
table. The specific viscosities (η – η0)/η 0 can now be calculated as (t – t0 )/t0 ; from
these values the reduced viscosities follow, and extrapolation to c = 0 (in this case
very simple, because the relation is linear) leads to:

[η] = 1.525 dl/g

c (g/dl) t (sec) ηsp (–) ηred (dl/g)

0 10
0.2 13.3 0.33 1.65
0.4 17.1 0.71 1.775
0.6 21.4 1.14 1.90

The intrinsic viscosity appears to depend on the molar mass according to the Mark-
Houwink relation:

[η] = k·Ma

in which k en a are constants for a given combination of polymer, solvent and

temperature. k is in the order of magnitude of 10-3 when [η] is expressed in dl/g; a is
mostly in between 0.5 and 1. The measurement of the molar mass is not an absolute
one; for each combination k and a have to be determined in an empirical way with
the help of absolute characterization methods such as light scattering. Values of k
and a are amply being supplied in handbooks.

The question is now, what kind of average molar mass is found from the intrinsic
viscosity and the Mark-Houwink equation. To answer this question, we imagine the
polymer to be split-up into a number of monodisperse fractions. Each of this
fractions brings about a viscosity increase, which for the ith fraction amounts to:

ηsp,i = (ηi – η0)/η 0 .

We assume that the total increase in viscosity is, simply, the sum of the contribution
of all fractions, while the concentration of the i th fraction is ci. Then

∑ ηsp‚i ∑ ( ηred,i·ci)
ηred = ∑ c = ∑ ci
Molecular composition 37

When the limit to c → 0 is taken, this becomes:

∑ [ η]i ·ci ci
[ η] = = ∑{[η]i · } = ∑ wi·[η]i
∑ ci ∑ ci
= wi
∑ ci

[η] = ∑ (wi ·k·Mi a) = k·∑ (wi ·Mi a ) ≡ k·(Mv )a

In this way the average Mv is defined as

Mv = [∑ wi·Mi a]1/a
— — — —
So, for a = 1, Mv = Mw; for usual values of a, such as 0.8, Mv is situated between Mn
— —
and Mw , but more closely to Mw . In practice the full procedure of measuring at
different concentrations and extrapolation to zero concentration, is often replaced by
a simpler standard method; for a given polymer in a given solvent, only one
concentration is taken. Optionally, a correction factor can be applied to obtain from
ηred the best estimate of [η].
Sometimes, from such a single-point measurement of the solution viscosity, an
arbitrary measure has been defined, which for a given polymer supplies a useful
measure for comparing levels of the molar mass. An example is the so-called k-value
for PVC; k values of 55, 60, 65 and 70 denote grades of increasing M. (see Qu. 2.40
and 2.41).
The melt index m.i. (or melt flow index m.f.i.) is a rough empirical measure of the
molar mass of some polymers. It is, again, based on a viscosity measurement, but
now not of a solution, but of the molten polymer. Through a standard capillary the
polymer moves under a standard pressure at a standard temperature. The number of
grammes transported in 10 minutes is defined as the melt index (in dg/min). The
method is frequently applied to e.g. PE and PP, and results in some measure for the
molar mass. The m.i. values range from 0.1 (high M) to 80 (low M) dg/min. In fact a

reciprocal melt viscosity is determined. In a first approximation this depends on Mw,

viz according to η (:) Mw3,4 (see § 5.3). In a certain sense this characterization of the
molar mass is functional, since the processability is one of the most important
criteria for the selection from a number of grades of the same type of polymer (see
§ 5.4).

Molar mass distribution MMD

The oldest method to determine the whole molar mass distribution is fractionation.
This method is based on the fact that in some solvents the short chains are better
soluble than the long ones. The polymer is precipitated onto a column, e.g. on small
38 From polymers to plastics

glass spheres, over which a mixture of solvent and non-solvent is passed while the
concentration of the solvent increases with time. First the smaller molecules are
taken away from the column, later on the bigger ones. The fractions are collected
and characterized as to their concentration and molar mass (e.g. by a viscosity
measurement). From this procedure the MMD can be derived in a number (e.g. 20)
of different fractions. This Baker-Williams method is very time-consuming:
fractionation and characterization of a sample takes, in total, about a week.
A more modern method to determine the MMD is GPC, gel permeation chromato-
graphy, also named size-exclusion chromatography, SEC. A polymer solution is
passed over a column with a porous structure. The residence time of the chains on
the column depends on the diameter of the coiled chain: smaller chains can migrate
through more pores (they can also enter into the smaller ones), and it takes a longer
time for them to pass along the column. The bigger ones cannot enter into any of the
side-pores and pass in the shortest time.
The concentration of the eluted solution is continuously measured as a function of
time (or of eluted volume), e.g. via its refractive index, in comparison to that of the
pure solvent.
This procedure results in a concentration - volume curve, from which, after previous
calibration, the molar mass distribution can be derived. Calibration can be carried out
with known monodisperse polymers, and is needed only once for a certain type of
polymer on a certain column. The measurement takes only a few hours. From the
measured MMD the various averages can be computed easily. It is also possible to
characterise the eluted polymer solution not only on concentration, but also on molar
mass, e.g. by laser light scattering. In this way the calibration can be avoided.

2.3. Chain regularity

Polymer chains are, in general, regularly built-up, but a few variations are possible.
We shall, successively, consider: arrangement of monomers, situation of side groups,
arrangement round a double bond, branching, and copolymer structure.

– monomer arrangement is, in principle, possible as “head-tail” or as a “head-head”

(or “tail-tail”) i.e. for a vinyl polymer –CH2 –CHR–:

–CH2 –CHR– CH2 –CHR– CH2 –CHR–

–CH2 –CHR– CHR– CH2 –CH2 –CHR–

Practically always the head-tail connection is formed. One of the exceptions is

polyvinylalcohol, in which between 1.5 and 2 % head-to-head sequences occur.
– situation of side groups; the chain – CH2 – CHR– can be regular or irregular:
Molecular composition 39


regular; all R’s are at one side: isotactic.


regular; the position of the R’s is alternating: syndiotactic.

irregular: atactic.
This can, more clearly, be seen in a three-dimensional representation, in view of
the actual valency angles (tetraedic, 109°); the –C–C– main chain is situated in
the plane of drawing (Figure 2.4), while the side groups protrude at the front side
or the backside. (syndiotactic configuration)



Figure 2.4. Syndiotactic chain.

With conventional polymerization processes, atactic chains are predominantly

formed; for the formation of isotactic and syndiotactic chains a special catalyst
system is required, e.g. Ziegler-Natta catalysts. Such a process is called:
stereospecific polymerization. It enables the manufacture of, i.a., technically
usable PP and also unbranched PE (see § 4.1). The newest development is the
metallocene katalyst; it enables the building-up of “chains-to-measure” with very
high degrees of chain regularity; also the manufacture of syndiotactic polystyrene
is technically possible in this way (see Qu. 2.47).

– Round a double bond the main chain can show two different configurations,
e.g. with polybutadiene and polyisoprene (Figure 2.5):



trans cis

Figure 2.5. Cis- and trans-configuration. E.g. polybutadiene and polyisoprene

40 From polymers to plastics

Cis-1,4 polyisoprene (natural rubber or synthetic isoprene rubber) and trans-1,4

polyisoprene (balata or guttah-percha) show strongly different properties.
Within a single chain cis- and trans configuration can also both occur in a random
sequence; the chain is than irregular (some kinds of polybutadiene). This has
consequences for the occurrence of strain-induced crystallization (see Chapter 4).

– Branching disturbs the chain regularity if the branches are situated at random
positions along the chain. In particular with PE a number of branching types
are present (see Figure 2.6):
a. strongly branched with irregular branches: low-density PE LDPE)
b.little branched: high-density PE, (HDPE)
c. strongly branched, but more regularly: linear-low-density PE (LLDPE)
The names reflect the effect of the branches on the crystallinity and thus on the
density and the stiffness.

Figuur 2.6. Three types of branching for PE.

– With copolymerization irregular chains are formed when the sequence of the
monomers is random::


This is called a random-copolymer. An example is styrene - butadiene rubber,

A more regular chain structure is the block copolymer:

–(A–A–––A–A)– (B–B–––B–B)–(A–A–––A–A)– (B–B–––B–B)––

examples are the thermoplastic elastomer SBS (see § 9.1.4) and the graft

and, in particular, the alternating copolymer:

Molecular composition 41

Strictly speaking, condensation polymers such as polyesters could be considered to

belong to this category. Also polyketone (“Carilon”) would be a strictly alternating
copolymer, if the regularity would not have been disturbed by a third comonomer,

2.4. Chain conformations

Polymer chains are hardly ever met in a totally stretched condition, but are, because
of their mobility and flexibility, present as coils. The coil can show all kinds of
shape, as dictated by chance; as a first approximation we may assume a spherical
shape. The diameter of this sphere is hard to define; the density in the coil decreases
from the centre to the outer boundary. The easiest criterion is the end-to-end
distance, r 0; this is a strongly fluctuating statistical quantity, of which we must,
therefore, try to define some kind of average. This average has some relation with the
effective coil diameter.

In order to estimate the end-to-end distance r0 we assume, as a first approximation,

that the chain segments can move freely with respect to each other in all directions.
The chain contains n segments, each with a length b0. The next simplification is, that
we consider the chain in two dimensions. We now have a simple “random walk”
(“drunk man’s walk”) problem. We situate one chain end in the origin of an x-y
coordinate system and we build the chain step by step with randomly chosen angles
ϕ (see Figure 2.7). The position of the other end is than given by

x = b0 ·cos ϕ1 + b0 ·cos ϕ2 + … + b0 ·cos ϕn

y = b0 ·sin ϕ1 + b0 ·sin ϕ2 + … + b0 ·sin ϕn

n n
x = b0 ·∑ cos ϕi; y = b0·∑ sin ϕi
i=1 i=1



Figure 2.7. ‘Random walk’.

42 From polymers to plastics

The end-to-end distance is r0 = 


r0 2 = x2 + y2 = b0 2 ·(∑ cos ϕi)2 + b0 2 ·(∑ sin ϕi)2

r0 2
= ∑i cos2 ϕi + ∑i∑ j cos ϕi·cos ϕj + ∑i sin2 ϕi + ∑i∑ j sin ϕi sin ϕj =
b0 2
=∑ (cos2 ϕi + sin2 ϕi) + ∑i∑ j cos(ϕi – ϕj) =

= n·1 + 0 = n

The second sum is 0 because ϕi – ϕj is distributed at random.

The statistical value (the average) of the square of the end-to-end distance is given

〈r0 2 〉 = n·b0 2 or √〈r02〉 = b0·√ n

Further calculation shows that for a three-dimensional random walk the same result
is obtained.

In a second approximation we have to introduce a fixed value for the angle between
two links of the chain; since, e.g., the angle between two valency bonds of the C-
atom is always θ = 109.5°. (Figure 2.8). For the time being, we suppose free rotation
around this angle. The approach is similar to the previous case (of course now three-
dimensional), and results in:

1 – cos θ
〈r0 2 〉 = n·b0 2 ·
1 + cos θ


Figure 2.8. Free rotation with fixed valency angle.

This formula is not valid when θ is close to 0° or 180°. It appears that the correction
Molecular composition 43

= 1 for θ = 90°
< 1 for θ < 90°
> 1 for θ > 90°
= 2 for θ = 109,5°
In a third approximation we also limit the freedom of rotation. This can be illustrated
with ethane CH3–CH3; when this molecule is rotated around the C–C bond, potential
barriers have to be overcome as a result of the interaction between the H-atoms (see
Figure 2.9). For PE a similar potential curve can be calculated; this has a more
complicated shape as also indicated in Figure 2.9.


–180 –120 –60 0 60 120 180 °


Figure 2.9. Potential curve for rotation

The hindering in rotation necessitates the introduction of an extra correction factor in

the formula for 〈r02 〉. This factor contains the quantity 〈cos ϕ〉, which is the average
cosine of the rotation angles. For free rotation 〈cos ϕ 〉 = 0 (all values for ϕ are
equally probable), but it differs from 0 when potential barriers are present. It can be
calculated from the potential curves. The formula becomes now

1 – cos θ 1 + 〈cos ϕ〉
〈r0 2 〉 = n·b0 2 ·
1 + cos θ 1 – 〈cos ϕ〉

A simplified formula is:

〈r0 2 〉 = n·b2

in which b is the effective bond distance, i.e. the length of the fictitious link, which
behaves without restrictions as a random walk. We maintain n as the number of
primary links (sometimes n is taken as the number of fictitious links), and we define
b/b0 as the characteristic chain stiffness. This quantity is, therefore = √  2 for a
–C–C- chain with unrestricted rotation; b/b0 > √ 2 with hindered rotation.
Still further approximations are required to account for the “excluded volume”: the
44 From polymers to plastics

chain parts cannot coincide. An important complication is met when the side groups
are big enough to hinder each other during chain rotations (see § 2.5). In all cases the
formula 〈r0 2 〉 = n·b2 remains valid, though b/b0 increases as a result of these effects.
Some estimated values:


(b/b0) 2.50 2.61 2.97 3.00 3.29

We can now ask the question: What is the magnitude of the coil density? How far is
the statistical coil, which we have considered so far, diluted? How many monomer
units are present in a unit volume? It is possible to calculate how the density, ρ,
depends on the distance to the centre of gravity and on the number of links n. It
appears that the volume fraction in the centre equals f0 = 1/√
 n, i.e.:
for n = 1,000 f0 = 0.03
for n = 1,0000 f0 = 0.01

in other words: the coils are, even in their centre, highly diluted!
A rough estimate can also be made without much theory, namely from 〈r0 2 〉 = n·b2 ,
for example for PE with M = 140,000. A chain contains 10,000 CH2 links, so n =
10,000. The length of a link is b0 = 0.154 nm, and we simply assume unrestricted
rotation, so that b = b0 ·√
 2 = 0.218 nm. We then find:
〈r0 2 〉 = 10,000·(0.218·10–9)2 m2

If we, loosely, think of a sphere with radius r0, then the volume of this sphere is:
V = 3 π ·〈r02 〉3/2 = 0.433·10–22 m3

The mass is m = 140.000 g / 6·1023 = 2.3·10–22 kg. The specific mass of the coil is:

ρ = m/V = 5.3 kg/m3

When we compare this value with the density of amorphous PE (ρ = 855 kg/m3 ),
then the polymer in the coil appears to be 160 times diluted. If we choose the value
for b/b0 as given above, then the result is a dilution factor of 890. Taking into
account that the formula f0 = 1/ √  n holds for the tightest packed centre, then the
results are in the same order of magnitude.
Coiled chains, are, therefore, strongly interwoven; in each volume element of the
size of a coil, parts of hundreds of separate chains are present!
The reasoning followed so far is based on the assumption that chains, in finding their
most probable coil dimensions, are not affected by their environment; in other
words, that no special interactions, either attractive or repulsive, with neighbouring
Molecular composition 45

molecules are present. This holds, therefore, for an undiluted polymer, in which a
chain does not undergo interactions from the surrounding chains other than those
from parts of the chain under consideration itself.
If we now look at polymer solutions, then there are three possibilities (see also Qu.
2.20 –2.23):
– A very good solvent; this means that there is an extra interaction between parts of
the chain and the molecules of the solvent, so that the chain segments prefer to
be surrounded by more solvent molecules. The coil will then become more
– A poor solvent; parts of the chain feel better in an environment consisting of
similar chain segments; the coil will become more compact.
– In between these cases we can define a so-called theta-solvent, which, as far as
interactions are concerned, behaves, at a certain temperature, like the polymer.

Our calculation of 〈r02 〉 is therefore, for polymer solutions, only valid for theta
conditions. We shall consider this case in somewhat more detail. We have seen that
for the determination of the viscosity average molar mass, the Mark-Houwink
equation is valid

[η] = k·Ma

in which, in general, a is in between 0.5 and 1. Moreover, it is known that under

theta conditions a = 0.5.

We have also seen that we have to extrapolate to concentration zero to obtain the
intrinsic viscosity[η] = lim ηred.
Considering the calculation of coil density, this is not surprising. A concentration of
1% in a solution seems small, but when the coil is more than 100 times diluted,
considerable mutual penetrations of the coils can be expected, and, therefore, strong
deviations from the behaviour of single, separate coils.
Only at very low concentrations can the coils be considered as separate spheres. In
that case Einstein’s relation for the viscosity of suspensions can be applied:

η = η0·(1 + 2.5·ϕ) ,

in which ϕ is the volume concentration and η0 the viscosity of the liquid. Then:

η – η0
ηsp = = 2.5·ϕ en ηred = 2.5·ϕ/c = 2.5/ρ

in which ρ is the specific mass of the coil, or:

46 From polymers to plastics

with M the molar mass and V the volume of the coil.

This, combined with the Mark-Houwink equation, yields:
[η] (:) Ma (:)
V (:) M1+a

If, as above, we express V as:

V (:) 〈r02 〉3/2

〈r0 2 〉 (:) number of links,so (:) M
V (:) M3/2 (:) M1+a

so: a = 0.5.
This (too) simple derivation confirms that for theta conditions: a = 0.5!
When another type of solvent is chosen, a higher value for a can be found, e.g. a = 1;
in this case:

〈r0 2 〉 (:) M4/3 instead of (:) M

or a lower a, e.g. 0.35; then 〈r0 2 〉 (:) M0,9.

Deviations of 〈r0 2 〉 (:) M can also be brought about by other causes, for example by
chain branching; as an extreme case we can think of a highly complex branched
structure in which the many branches are again branched, so that it can hardly be
defined as a chain. The conformation statistics is then no longer applicable, and for
such polymers a value of a near to 0 has even be found; this means that V (:) M, just
as with small molecules. Viscosimetric determination of M is then no longer
applicable, since [η] is independent of M.
A deviation in the other direction occurs with super-stiff chains, in the extreme case
with rigid rods. The end-to-end distance then equals the whole chain length, so:

〈r0 2 〉 (:) M2

which means a = 2.
This value is, therefore, mentioned as the theoretical maximum of a; it remains an
open question what the significance of V is in such a case, thinking of our spheres!
For liquid-crystalline polymers values round a = 1.5 are indeed being found.
Molecular composition 47

2.5. Chain flexibility

The flexibility of chains is governed by the freedom of rotation of the main chain,
and also by the effect of side groups. First we consider the main chain. In the
foregoing we have already seen that the possibilities for rotation of the main chain
are restricted by potential barriers. Some examples for simple compounds are:

CH3 –(CH3 ) 2.8 kcal/mol

CH3 –(CH2 –CH3 ) 3.3
CH3 –(CH.CH3– CH3 ) 3.9
CH3 –(OH) 1.0
CH3 –(O– CH3 ) 2.7
CH3 –(COH) 1.0
CH3 –(C≡C) 0.5

Double (C=C) and triple (C≡C) bonds cannot rotate; the ease of rotation of the
neighbouring bonds is, however, increased.

The series shown below represents a number of types of main chains with decreasing
– C– O – C– e.g. POM, polyoxymethylene
– C–N – C– e.g.PA, polyamide, nylon
–C= C– C– e.g.BR, butadienerubber
– C– C– C– e.g.PE, polyethene
– –C– e.g.PC, polycarbonate
– – PI, polyimide

Side groups mainly affect chain flexibility when the main chain is flexible. An
example is steric hindrance with (see Figure 2.10)




0.252 nm

Figuur 2.10. Steric hindrance.

48 From polymers to plastics

X = H: R = 0.09 nm, little hindrance

X = F: R = 0.14 nm, some hindrance
X = Cl: R = 0.18 nm, considerable hindrance
X = CH3 : R = 0.20 nm, considerable hindrance

The situation sketched above is not quite exact, since the substitution of a side group
also causes a slight change of θ.
When the sterical hindrance is considerable, no zig-zag conformation of the chain is
possible; the number of conformations is then drastically reduced.
The effect of a restriction in rotation is an increase of b/b0, and, therefore, of the

effective coil diameter 〈r0 2 〉.
The chain flexibility has, a.o., a large influence on the glass-rubber transition
temperature, and (if applicable) on the melting point.

2.6. Chain interactions

Between parts of the chains interaction forces are active, the secondary binding
forces, which are much smaller than the primary chemical bonds by which the
chains are held together. Several kinds of secondary binding forces exist:

– dipole forces; here the distribution of electrical charges within an atom or group
of atoms is asymmetric with respect to the centre of gravity (see Figure 2.11).
Polar groups form permanent dipoles, which align in the electric field of their
neighbours; this results in an electrical attraction force. The interaction is
proportional to 1/r3 (r = distance), and is slightly dependent on the temperature as
a result of the thermal motion of the dipoles. A measure for the interaction force
is the dipole moment, the value of which is, for a few groups, given in the table

++ –
+ – + – + – –
+ –
+ –

permanent dipole induced dipole

Figuur 2.11. Permanent and induced dipoles.

– induced dipole; this is formed when a neutral group is being polarized by

induction from a neighbouring permanent dipole. The attraction force is proport-
ional to 1/r6 and is slightly temperature dependent.
Molecular composition 49

CH3 –CH2 –X

X dipolemoment X dipolemoment
–CH3 0 –COOH 1.73
–O– CH3 1.2 –COOCH3 1.76
–NH2 1.2 –COCH3 2.78
–OH 1.69 –NO2 3.68
–Cl 2.05 –CN 4.0
– dispersion forces originate from random fluctuations in the charge distribution of
an apolar group as a result of the revolution of electrons. A fluctuating dipole
induces another one; the net result is an attraction force. Dispersion forces, also
called London-Van der Waals forces, are independent of temperature and
proportional to 1/r6 They form the most important class of interaction forces, as
for instance in hydrocarbons.
– hydrogen bridges occur between chains in e.g. the following situations (Figure


Figuur 2.12. Hydrogen bridges.

The H-valency is being shared with the neighbouring atom, e.g.:


This leads to relatively strong bonds, as in nylon and cellulose. The hydrogen
bonds are also responsible for the strong link between water molecules; that is
why water, as a very small molecule, has a high boiling point!
Consequences of chain interactions for the behaviour of a polymer are, just as with
chain flexibility, mainly the height of the glass-rubber transition and the melting

2.7. Cross-linking
The most important characteristic of a network is the tightness, the spacing between
cross-links. Some quantities frequently used will be elucidated in a simple schematic
Suppose that we have two chains, each of 20 links, which are, at equal distances,
connected by cross-links, as illustrated in Figure 2.13.
50 From polymers to plastics

Figure 2.13. Cross-links.

The cross-link density is defined as:

number of units connected 6
ρ= total number of units
= 0.15

The quantity ρ is between 0 and 1; ρ is low for vulcanized rubbers, namely 0.01 –
0.02, and much higher for thermosets.
We further define:

M0 = molar mass of the monomer (here e.g. M = 50)

Mc = molar mass between cross-links (here 5 × 50 = 250)
M = molar mass of the whole chain (here 20 × 50 = 1000)
ν = a number of chain elements, separated by cross-links (here ν = 8)
νe = effective number of chain parts, contributing to the coherence and
the elasticity of the network (here νe = 4).

νe is related to ν, M and Mc according to

νe = ν·[1 – 2· M ]

since the chain is divided into n = M/M c parts, two of which are always loose, so
that the effective fraction is (n – 2)/n = 1 – 2/n.

In our case νe = 8(1 – 2 × 250/1000) = 4.

In this example only one half of the chain parts are effectively contributing to the
network properties; the remaining part, the loose ends, are not being deformed when
the material is strained.

The various parameters can be determined from measurements of the modulus of

elasticity (see § 5.1) and from swelling in solvents.
In most cases, cross-linking is effected by the formation of chemical bonds between
the chains; the network is then stable, such as with:
– sulphur bridges with rubbers,
– polystyrene chains with unsaturated polyesters,
– bivalent with three- (or more-)valent reacting components,
– cross-linking by irradiation; via radical formation a chemical link between chains
is formed.
– other cross-linking agents such as peroxides.
Molecular composition 51

Also physical cross-linking occurs with:

– crystallization (see § 4.5 and § 7.2.1),
– domain formation in block copolymers, “self-vulcanizing rubber” or “thermo-
plastic elastomer” (see § 9.1.4),
– temporary entanglements, causing rubbery behaviour in unvulcanized rubbers
and in molten thermoplastics.

Glassy state and glass-rubber

3.1. Glassy state

In the usual schedule “solid - liquid” the solid is crystalline and passes into the liquid
state at the melting point, Tm. This transition is, in nearly all cases, accompanied by
an increase in volume (one of the exceptions is water!), and with an increase in the
heat content (enthalpy), the heat of melting.
The jump in volume is illustrated in Figure 3.1; the slope of the line FC is the
thermal expansion coefficient of the crystalline phase; at the melting point the
volume jumps from C to B, and the higher slope of BA denotes the expansion
coefficient of the liquid phase.
Some substances are, however, not able to crystallize, for instance normal glass, as a
result of a too irregular molecular structure. When such a substance is cooled down
from the liquid state, and follows the line AB, then from B to D it still remains a
fluid, which solidifies at D without showing a jump in volume. The line then
continues as DE, with about the same slope as CF; the matter is, however, not in a
crystalline condition, but in an unordered, amorphous, glassy state, and has,
therefore, a greater volume.

V fluid
glass D
F crystalline

Tg Tm T

Figure 3.1. Volume as a function of temperature.

The transition at D is called the glass transition, occurring at the glass transition
temperature, T g. It follows that Tg is always lower than the melting point, T m . It is
very important to distinguish very carefully between Tg and Tm !!
Glassy state en glass-rubber transition 53

Polymers are sometimes wholly amorphous and non-crystallizable; they then follow
the line ABDE. However, when such a polymer is heated up to above its Tg it is not
immediately transferred into a liquid state, but first into a rubbery state, which, upon
further heating, gradually passes into a fluid. Tg is, therefore, called the glass-rubber
transition temperature.
A more realistic representation of the phases is given by the stiffness of the material,
which falls down to zero when the liquid state is reached. This is illustrated in Figure
3.2; for a low-molecular matter the E-modulus (a measure of the stiffness) decreases
to zero at the glass transition temperature (though more gradually than for a
crystalline substance at Tm). A polymer shows, above Tg , after a decrease in E by a
factor of 1,000 to 10,000, a rubbery region, which, on the temperature scale, is
longer as the chains are longer.

10 9

10 8
10 7

10 6

10 5
Tg Tm T

Figure 3.2. Rubbery region with polymers.

Sometimes crystallization is possible; in the solid state polymers are, however, never
totally crystalline, but are still partly amorphous. The amorphous part may be above
or below Tg , i.e. in the rubbery or in the glassy condition.
The nature and the properties of polymers are, therefore, governed by two different
schedules, which are sometimes superimposed:
1. glass — rubber — liquid
2. crystalline — liquid.

First of all we shall consider the glassy condition.

In the glassy state the molecular structure is disorderly, and comparable to that of a
liquid. This is clearly demonstrated by X-ray diffraction patterns, in which only a
diffuse ring is visible, which indicates some short-distance order; in contrast to to the
sharp reflections found with crystals as a result of long-distance order.
The disordered state occupies a larger volume than a crystal, which explains the
distance between the lines ED and FC in Figure 3.1, the so-called free volume.
54 From polymers to plastics

Below Tg the free volume, V f , is about constant; above Tg it increases strongly with
increasing temperature.
The free volume in the glassy state allows some minor movements for small chain
parts or for side groups. These movements are possible from a certain temperature;
they are apparent from a relatively small decrease in stiffness at that temperature.
Because such a decrease is much less than in the glass-rubber transition, the
temperature at which it occurs is called a secondary glass transition. Besides the
modulus, E, also the mechanical damping is a good criterion for the transitions. The
damping is defined as the tangent of the loss angle in vibration experiments, or, also,
as the rate of damping in a free vibration (see § 6.2). The damping, tan δ, shows a
strong maximum at Tg, and less pronounced maxima at the secondary glass
transitions Tsec (see Figure 3.3.).
A secondary glass transition is, in general, important for the impact strength of a
polymer; it creates the possibility to dissipate energy in situations of shock loading,
so that the polymer is less brittle.

log E

tan δ

Tsec Tg T

Figure 3.3. A secondary glass transition.

Besides by motions of chain parts or side groups (see § 3.4), a secondary transition
can also result from the presence of a second polymer, which has been added in a
small quantity by blending, or which has, as “tails”, be copolymerized with the main
polymer. Such a second polymer then has its glass transition at Tsec. Both cases occur
when the impact strength of a polymer has been improved, either by blending with
some rubber, or by copolymerization. A similar situation is found with block
copolymers. In § 3.5 these cases are dealt with in more detail.

3.2. Molecular picture

Before looking closer at the glass rubber transition, we shall try to visualize what
happens on a molecular scale when a polymer is heated, starting from zero Kelvin.
At T = 0 K the chains are at absolute rest. No thermal motions occur; everything is
Glassy state en glass-rubber transition 55

completely frozen-in. The chain parts are bound together by secondary interaction
forces, as discussed in § 2.6.
When the temperature is increased from T = 0, the thermal motion increases and,
gradually, short parts of the chain or side groups may obtain some mobility, which,
within the very restricted free volume, gives rise to small changes in conformation.
Whether or not this occurs, is a matter of competition between the thermal energy of
a group (kT) and its interaction with neighbouring groups.
The interaction can be expressed as a potential barrier which has to be overcome in
order to realize a change in position. Since the thermal energy is statistically
distributed, a fraction of the groups will be able to overcome the barrier; this fraction
strongly increases when the temperature is raised. Hence, there is a characteristic
transition temperature for such a specific molecular motion.
This is illustrated in Figure 3.4. The potential barrier for passing from state 1 to state
2, is ∆U; the distribution of the thermal energy has been sketched schematically for
three temperature levels. At T1 only a small number of groups have enough energy to
overcome the barrier, at T 2 much more, and at T3 nearly all. Thus T2 is, for that
displacement, the transition temperature. For T < T 2 the mobility is frozen-in; for
T > T2 the motion is free.

∆U T2 > T1

T3 > T2
1 2
∆U U

Figure 3.4. Potential barrier for changing position.

We should realize, that this is always a matter of probability, and that the time, i.e.
the patience we have to wait for a change, plays an important role. Therefore, we
should consider the frequency at which jumps from state 1 to state 2 occur. This
frequency, ν, can be expressed by the the Arrhenius equation:

ν = ν0 exp(–∆U/kT)

In this equation ν0 is the natural frequency of the vibrations about the equilibrium
position. The jump frequency governs the time scale, τ , at which the transition
occurs; τ is namely inversely proportional to ν:
56 From polymers to plastics

τ = A exp(∆U/kT) .

This equation provides a fundamental relationship between the effects of time and
temperature on a transition mechanism. Time and temperature appear to be
equivalent in their effect on the behaviour. At a fixed time scale, τ 1 , e.g. 1 sec, the
transition temperature, T1 , is proportional to the energy of activation, ∆U. The
transition temperature can be expressed in the time scale by:

ln τ = c +
Consequently. for low-temperature transitions with a low energy of activation, such
as secondary glass transitions, the transition temperature is stronger time dependent
than, e.g. the main glass-rubber transition.
An example of a secondary transition is the side group rotation -CO-O-CH3 with
PMMA, which is set free at about 25 °C on a time scale round 1 sec. PVC shows a
weak transition at –30 °C as a result of freedom of rotation for the Cl atoms. A very
elegant illustration is given by polycyclohexylmethacrylate in comparison with
polyphenylmethacrylate (Heijboer). The polymers are nearly identical; the difference
is that the first has a saturated ring as a side group, and the second an unsaturated
The saturated ring is able to undergo a “flipping motion” (chair-chair transition) at
–80 °C, which, in contrast to the rigid ring in PCHMA, results in a strong secondary
transition (see Figure 3.5). With further increase of the temperature, bigger parts of
the chain can move freely. About Tg the mobility increases drastically, so that coils
can freely be transformed to other conformations, without being hindered by
interactions: the polymer shows rubbery behaviour. This is the case for free mobility
of chain parts with, on average, 20 to 70 monomeric units. Chain entanglements
prevent whole chains to move relatively to each other.

3.3. Thermodynamics of the glass-rubber transition

To consider the nature of the glass rubber transition on a thermodynamic basis, we
should first compare this transition with melting. The melting point is a first-order
transition; Tg could be mentioned a second-order transition.
The quantity G, the (Gibbs) free enthalpy, plays a predominant role in the
thermodynamic treatment of transitions.

G = U – TS + pV = H – TS = F + pV

in which
Glassy state en glass-rubber transition 57

CH 3

CH 3


Figure 3.5. Effect of the nature of the side group on the secondary glass transition (J.
Heijboer, ‘Mechanical properties of glassy polymers containing saturated rings, dissertation
RU-Leiden, Waltman, Delft, 1972).

U = internal energy, e.g. as a result of the attraction forces between molecules

T = absolute temperature,
S = entropy, a measure for the disorder in the system,
p = pressure,
V = volume,
H = U + pV = enthalpy or heat content
= internal energy + work exerted on the environment,
F = U – TS = (Helmholtz) free energy.
With each type of transition, ∆G = 0, in other words: the G(T) curves for both phases
intersect, and slightly below and above the transition temperature the free enthalpies
are equal. The various derivatives of the free enthalpy may, however show
discontinuities. With a first-order transition such as melting, this is the case with with
the first derivatives like V and S and also with H.
58 From polymers to plastics

[∂G ] =V
∂p T

With melting the volume differs at both sides of the melting point; a jump ∆V in
volume occurs, which is a jump in the first derivative of G after p at constant T. Also
a jump ∆H occurs in the enthalpy (the heat of melting), as well as a jump ∆S in
The glass-rubber transition, on the contrary, does not show jumps in V, S and H (no
volume change, no discontinuous change in the state of order and no heat effects).
However, jumps occur in the derivatives of these quantities, such as thermal
expansion coefficient, specific heat and compressibility. Some examples:

[∂G ] = V,
∂p T
Vglas = Vrubber, ∆V = 0

[∂p∂ G∂T ] = [∂V

] = αV,
∂T p
αglass ≠ αrubber

the coefficient of expansion shows a jump ∆α .

So there are jumps in the second derivatives of the free enthalpy G, and, for that
reason, the glass-rubber transition might be denoted as a second-order transition.
However, this appears not to be the case; the dependence of the various parameters at
a real second-order transition strongly differs from what happens at Tg , which will
not be further shown here.
A second argument is the time dependency: the transition from a supercooled liquid
or rubber to the glassy state is clearly dependent on the rate of cooling; the line
representing the liquid behaviour continues further to lower temperatures as the rate
of cooling is lower; the glass transition temperature is then lower and the specific
volume reaches a lower value (see Figure 3.6).
A thermodynamic equilibrium is, therefore, not reached; if such an equilibrium
would exist it would be at a much lower temperature. On the basis of theoretical
calculations it has been supposed that a real second-order transition could occur at a
temperature which is 50 to 60 K below the observed value of Tg , and which would be
reached after extremely low rates of cooling. Estimations on the basis of the
empirically found relation between Tg and the rate of cooling indicate that the
required cooling time would be of the order of 10 17 years!
Experiments have also been carried out on polystyrene with various amounts of
plasticizers; extrapolation to zero plasticizer content provided an insight into the
course of the volume-temperature relation below Tg. Even at (T g – 70 K) no
indication of a transition was found.
Glassy state en glass-rubber transition 59

It thus seems that the glass transition, even with infinitely low cooling rate, is not a
real thermodynamic transition, but is only governed by kinetics, as a freezing-in

cooling A


slow cooling

Figure 3.6. Effect of the time scale on the glass transition.

A closely related phenomenon is the volume retardation in the glassy condition.

When, after rapid cooling, point A has been reached (Figure 3.6), the volume will, at
a constant temperature, gradually decrease to e.g. A′. Possibilities for motion of
chain parts, though on a very limited scale, cause the free volume to decrease
gradually, though, as we now know, an equilibrium will never be reached. The rate
of volume retardation is, as a matter of fact, smaller as the temperature is lower.
This phenomenon, also called physical ageing, is not very significant in terms of
volume, but, as we shall see later on (§ 7.2.4.), it is very important for the creep
behaviour of polymers.

3.4. Factors governing Tg

The height of the glass-rubber transition temperature is, in the first instance,
governed by the competition between thermal motion and the attraction forces
between the chains.
The thermal motion is dependent on the freedom of the chain to undergo changes in
conformation. When this freedom is higher, the chain is subjected to a stronger
thermal motion than a chain which, e.g. as a result of hindrance in rotation, is more
rigid. The chain stiffness, as discussed in § 2.4, plays, therefore, an important role.
The primary criteria are thus:
– chain flexibility
– chain interactions.
We shall consider some examples of both criteria. In doing so, besides amorphous,
also crystalline polymers will be taken into account, since the latter always contain
60 From polymers to plastics

an amorphous fraction, which is subject to a glass transition.

3.4.1. Chain flexibility

Higher chain stiffness results from a smaller number of possible chain
conformations; this can be caused by:
– greater stiffness of the main chain,
– bigger side groups,
– cross links.

First a few examples of chain stiffness differences in the main chain; the examples
are not very exact, because side groups may also play a role as to flexibility or
interactions, but they represent a global trend.
–[CH2 –CH2 ]– PE, polyethylene, Tg = –120 °C

–[CH2 –CH2 –O– CO– –CO– O–]– PETP, polyethyleneterephtalate, Tg = 70


–[S– –]– PPS, polyphenylenesulfide, Tg = 85 °C

–[O– – C– –O– CO]– PC, polycarbonate, Tg = 150 °C
–[N N– –O– –]– PI, polyimide, Tg > 400 °C.

Now we look at some examples of the effect of side groups on the chain flexibility:

–[CH2 –CH2 ]– PE, polyethylene, Tg = –120 °C

–[CH2 –CH]– PP, polypropylene, Tg = –15 °C
–[CH2 –CH]– PS, polystyrene, Tg = 100 °C

–[CH2 –CH]– PVK, polyvinylcarbazole, Tg = 210 °C


The increasing size of the side group effects a decrease of chain flexibility and an
increase of Tg .
Glassy state en glass-rubber transition 61

Cross-links limit the mobility of macromolecular chains, and thus give rise to an
increase in T g . With a very low concentration of cross-links this effect is hardly
detectable. Experiments with the system polystyrene + divinyl benzene showed that
at a concentration of 0.4 % of DVB and higher a detectable increase of Tg occurred.
With rubbers, vulcanized with sulphur, the detection limit for an increase of T g
appeared to be at 1 to 2 % of sulphur, so that for a technically vulcanized rubber the
level of Tg is hardly influenced by the vulcanization process.
Higher degrees of cross-linking have a a significant effect; for an unsaturated
polyester resin with short and stiff cross-links, Tg appeared to be increased by 21 °C
when the number of atoms between cross-links amounted to 50, while with 24 atoms
between cross-links the increase was 93 °C. So, with thermosets the cross-link
density has a large effect on T g , and, consequently, on the usability at higher

3.4.2. Chain interactions

In § 2.6 we have looked at the types of interaction forces which can occur between
chains. The strongest of these are the dipole forces. Their effect on Tg is illustrated
by the series PP, PVC and PAN, in which the chain mobility hardly varies because
the side groups are of about equal size, but in which, in the order of sequence
mentioned, the dipole interactions increase.
–[CH2 –CH]– PP, polypropylene, Tg = –15 °C
–[CH2 –CH]– PVC, polyvinylchloride, Tg = 90 °C
–[CH2 –CH]– PAN, polyacrylonitrile, Tg = 120 °C
(see also Qu. 3.7).
Interactions can be decreased by increasing the distance between the chains, for
instance with long side chains, which lower Tg. This effect appears to be greater than
the increase of chain stiffness, as shown in the examples below; with PMP the more
compact structure of the side group gives rise to a different balance of these
counteracting effects.

–[C–C]– PP, polypropylene, Tg = –15 °C

62 From polymers to plastics

–[C–C]– PB, polybutylene-1, Tg = –25 °C

–[C–C]– polyhexene-1 , Tg = –50 °C
–[C–C]– polymethylpentene, Tg = 30 °C

Increase of the distance between chains and shielding of interactions play a very
important role in plasticizing of, e.g., PVC:
PVC Tg = 90 °C

PVC + 30% dioctylphtalate (DOP) Tg = 0 °C

3.4.3. Chain length

An empirical relation for the effect of the molar mass M on Tg is:

Tg = Tg∞ – K/M

in which Tg∞ is the value of Tg for infinitely long chains and K is a constant which is,
e.g., about 105. Consequently, for practically used polymers Tg can, at the most, vary
by a few degrees C. It can be argued that the effect of M on Tg is related to the
number of chain ends. Therefore, In the above formula, M denotes the number-

average molar mass, Mn . For branched polymers we meet the complication that each
chain has more than two chain ends; a large number of long branches decreases the
glass transition temperature.

3.4.4. Effect of time scale

The effect of time scale on T g has been discussed before; with very rapid cooling
values can be observed which are 5° or even 10 °C higher than with very slow
Glassy state en glass-rubber transition 63

3.5. Glass-rubber transition of blends and copolymers

Polymer blends can be subdivided into two kinds: those of compatible and those of
incompatible polymers. Real compatibility is an exception (see § 9.1); an example is
PS with PPE (polyphenylene ether, also called PPO, polyphenylene oxide). These
two polymers can be blended with each other on such a small scale that it really
looks like molecular miscibility. This blend shows, therefore, only one single glass
Most polymer blends are, however, composed of incompatible polymers; the blend is
then a dispersion rather than a real mixture; both components retain their own
individuality. In such a case, each of the two polymers shows its own glass
A schematic picture of both cases is given in Figure 3.7 .
log E blends log E dispersions


Figure 3.7. Glass-rubber transitions with real mixtures (left) and with dispersions (right).

With copolymers an analogous situation is met as with blends: Here also the
components may be molecularly compatible or not. When the copolymer is
“random”, i.e. the building blocks are arranged in a random sequence, then the
copolymer is, as a matter of fact, homogeneous on a molecular scale, even if the
components are incompatible. An example is SBR, a copolymer of styrene and
butadiene. It shows a single glass transition at about –65 °C, which is roughly in
accordance with the styrene content (23%) and the Tg ’s of polybutadiene (–95 °C)
and of polystyrene (90 °C).
The situation is quite different with block copolymers. As an example we again take
a copolymer of styrene and butadiene, but now as a three-block copolymer, SBS.
The incompatibility of polystyrene and polybutadiene now results in a phase
separation, which is enabled by the circumstance that the blocks can “live their own
life”. The polystyrene chain ends clog together into PS domains, which lie embedded
in a polybutadiene matrix. These glassy domains act as physical cross-links, so that
the polymer has the nature of a thermoplastic rubber. The glass-rubber transitions of
PS and BR both remain present; in between these two temperatures the polymer is in
a, somewhat stiffened, rubbery condition (see Figure 3.8). This behaviour is dealt
64 From polymers to plastics

with in more detail in § 9.1.4.

log E



Figure 3.8. Glass-rubber transitions of two kinds of copolymer.

3.6. Methods of measuring T g

Most methods to measure Tg are, respectively, based on the jump in specific heat and
on the sharp decrease of the modulus of elasticity, accompanied by a maximum in
the damping.
The specific heat as a function of temperature is measured by calorimetry, usually
with DSC (differential scanning calorimetry). At Tg a jump to a higher level is found.
The stiffness modulus is, in most cases, measured in dynamical - mechanical
experiments, for instance with a torsion pendulum, on a time scale of a few seconds.
This experiment results in the shear modulus, G (which is related to Young’s
modulus, E), while the damping shows a strong maximum at Tg.
Finally, the volume can be measured as a function of the temperature with the aid of
a dilatometer. At Tg a change in slope is detected, signifying a jump in the thermal
expansion coefficient.

L.C.E. Struik, ‘Physical aging in amorphous polymers and other materials’, thesis TH-Delft,
TNO, Delft 1977.

J. Heijboer, ‘Mechanical properties of glassy polymers containing saturated rings, thesis RU-
Leiden, Waltman, Delft, 1972.

Crystalline polymers

4.1. Conditions for crystallization

A primary condition (although not sufficient) for crystallizability is a regular chain
structure, which enables the formation of crystals in a lattice structure. This condition
can be met in several ways:
– when the side groups on both sides of an atom in the main chain are equal, so that
the chain is symmetrical. Examples of such chains are:

polyethylene PE –C–C–
polytetrafluorethylene PTFE – C– C–
H Cl
polyvinylidenechloride PVDC – C – C–
H Cl
polyisobutylene PIB – C–C–
polyoxymethylene POM – C –O–
polyphenyleneether PPE – –O–
polyphenylenesulfide (PPS) – –S–

polysulphone PSU – – C–– –O– – S– –O–

– when the side groups are small enough to be fitted into a crystal lattice.

66 From polymers to plastics

polyvinylalcohol PVAL;

– C – C–
the OH-group is small enough, to allow, with some distortion, crystallization into
a polyethylene-like crystal lattice.
polychlorotrifluoro-ethylene PCTFE;

– C– C–
F Cl
the Cl atom is not so much bigger than the F atom that it would prevent
crystallization; with PVC, however, the sizes of the H and Cl atoms are too much
different to allow crystallization:.

– C – C–
H Cl

– only one side group to an atom in the main chain:

polyethyleneterephtalate PETP,

|| ||
–C – –C –O– C – C –O–
polycarbonate PC,

| O
–O– – C– –O– C –
polyamides PA,

– – – C – C – N–C – C – C– – –
– regular arrangement of side groups (stereospecific)
Examples: (see also § 2.3)
polypropylene (isotactic) PP
| | | |
Crystalline polymers 67

polybutylene (isotactic) PB

| | | |
| | | |

Non-stereospecific vinyl polymers may yet contain enough regular sequences,

mostly syndiotactic ones, to allow a small amount of crystallization. Examples

polyvinylchloride PVC –CH2 –CHCl– (very small), and

polyacrylonitrile PAN –CH2 –CHCN–

– With unsaturated main chains the configuration around a double bond is of

importance. Only predominantly cis- or trans-chains are crystallizable. Here also
stereospecific polymerization is required. Some examples (see also § 2.3):

cis-1.4-polybutadiene, BR:

–C C–
H C=

cis-1.4-polyisoprene, IR, also natural rubber, NR:

–C C–
H C= C
H CH H 3

– With all these examples it should be remarked, that not all polymers which are, in
principle, crystallizable, actually crystallize spontaneously. The rate of
crystallization may be too low, such as with PPE, PC, PIB, BR and IR. In
particular for the rubbers (the last three mentioned) it can be said that they
crystallize when strongly strained, in other words, when the chains are highly
oriented (see also Qu. 4.1 to 4.3).

4.2. The melting point

At the melting point, T m , the solid and the liquid phases are in thermodynamic
equilibrium with each other. In § 3.3 we have, already, seen that the free enthalpies
are then equal:

G2 = G1 of H2 – TS2 = H1– TS1

68 From polymers to plastics

This leads to the following simple expression for the melting temperature:

H2 – H1 ∆H
Tm = S – S = ∆S
2 1

For the glass rubber transition a similar expression does not exist, since then ∆H and
∆S are both zero.
∆H is the enthalpy difference between crystal and liquid, the heat of melting. ∆H is
mainly governed by the difference in interaction forces in the solid and the liquid
condition; it is the energy required to disturb the crystal lattice.
∆S is the entropy difference between crystal and liquid and is related to the increase
in disorder when a crystal melts, or with the probabilities of the chain conditions in
both phases. This follows from:

S = k ln W (W = probability)
∆S = k ln

∆S is, therefore, controlled by the increase in the number of conformations of the

chains upon melting, so, a.o. by the chain flexibility.
Consequently, we meet the same criteria as with the glass rubber transition, namely
interactions and flexibility, but now on a quantitative thermodynamic basis. The
examples which can be given for the effects of interactions and flexibility on the
melting point are, therefore, similar to those for the glass transition:
Higher interaction forces give rise to a higher Tm:

PP Tm = 170 °C

PVC Tm = 280 °C (syndiotactic)

PAN Tm = 320 °C (compare with § 3.4)

Increase of chain stiffness results in a lower ∆S and thus in a higher Tm:

PE Tm = 135 °C

PS Tm = 240 °C (isotactic)

PTFE Tm = 325 °C (compare with § 3.4).

It is not always as easy to draw such conclusions. For the effect of chain flexibility
Crystalline polymers 69

on Tm, sometimes the example is used of polyethylene oxide, PEO, – CH2 –CH2 –O–
(Tm = 67 °C) in comparison with polyethylene, PE, –CH2 –CH2 – (T m = 135 °C),
with the argument that the C–O rotation is easier than C–C.
If we, however, consider a longer series of the type –(CH2 )n –O–, then we are in
trouble with our simple reasoning:
n = 1, polyoxymethylene, Tm = 181 °C
2, polyethyleneoxide, 67 °C
3, polytrimethyleneoxide, 35 °C
4, polytetrahydrofurane, 37 °C
∞, polyethylene, 135 °C.
There are, apparently, other effects which play a role, possibly the polarity of the
–O– segment in the chain and the topology of the chains in the crystal lattice; the
magnitude of the interaction forces does not only depend on the nature of the
attracting groups, but also on their distance. In this respect a comparison between the
crystalline densities of PTHF (1.1 g/cm3) and POM (1.5 g/cm3 ) is instructive; the
POM crystal is considerably denser, so that the real interactions play a much bigger
role in their effect on the melting point.
A second example: when we compare the Tm’s of a polyamide, e.g. PA-6.6 (260 °C)
and polyethylene (135 °C), then we might, superficially, conclude that the higher
value for PA is caused by a higher ∆H as a result of the strong hydrogen bridges
active in PA:

CH 2 CH 2 C N CH 2 CH 2

From experiments it appears, however, that the heat of melting, ∆H, of nylon is
lower than that of PE! From T m = ∆H/∆S it follows, that ∆S for PA must be
considerably lower than for PE. The explanation of this discrepancy is, that in PA the
H-bridges are still existent in the melt; they are not totally destroyed upon melting,
so that ∆H is lower than expected). Moreover, the H-bridges still present in the melt,
limit the chain mobility and thus the number of conformations, so that ∆S is
considerably reduced.
The effects of the hydrogen bridges indeed decrease with decreasing number of -CO-
NH- groups in the chain, so with increasing n in PA-n. With high n-values the
melting point approaches the value for PE (Figure 4.1, see also Qu. 4.16 and 4.17)).
70 From polymers to plastics

– The melting point is slightly influenced by the chain length according to the

1 b
=a+ (Figure 4.2)
Tm M

For commonly applied polymers this effect is only small. (See also Qu. 4.6.)



Figure 4.1. Melting points of polyamides.


Figure 4.2. Effect of molar mass on melting point.

– Chain orientation influences the melting point, which is understandable from the
equation Tm = ∆H/∆S. When the polymer is stretched in the rubbery phase, its
entropy is already so much decreased (see Figure 4.3 and, later on, in § 5.1), that
∆S has decreased. For natural rubber (cis-1.4-polyisoprene) a relation between Tm
and the strain has been established experimentally as shown schematically in
Figure 4.4.
– Irregularities in the chain decrease the melting point because the crystal becomes
less perfect. An example is polyethylene, PE; when the chains are not branched
the melting point is 135 °C; with three branches per 100 chain segments it is only
110 °C. An imperfect crystal has a lower ∆H.
– Disturbances in a crystal can be brought about by the above mentioned or by
other causes. These disturbances are arbitrarily distributed over the crystallites, so
that a spread in ∆H occurs, resulting in a spread in Tm , a melting region, which
has, as an upper limit, the melting point of the “best” crystal.
Crystalline polymers 71

∆S ( chain

∆S (isotropic

Figure 4.3. Entropy effects with straining and with melting.

80° T

0 600%

Figure 4.4. Melting point of natural rubber as a function of strain.

– The melting point is also somewhat dependent on the crystal size; in particular,
when the crystals are very small, T m is lower. This is, i.a., caused by the
restrictions in chain conformations near the crystal surface. A spread in crystal
size, therefore, also contributes to the presence of a melting region. This is clearly
demonstrated by the (new) PE grades with very low density (copolymers of
ethylene with e.g. 1-octene); the crystallites are so small that for a grade with d
=.86 the melting point is only 30 °C!
– Another striking example of melting point depression is found with a new kind of
isotactic polypropylene, made with the new “metallocene” katalyst. It is being
stated that everything you wish can be made with this new technique. Such as,
also, a PP with a very high degree of isotacticity, viz. 97% You would, therefore,
expect a higher melting point, but, unfortunately, it appears to be only 148 °C
instead of the usual value of 165 °C. What is the reason? Well, because “normal”
PP, made in the Ziegler-Natta process is indeed isotactic, but it contains a number
of chains which are not isotactic; these are mixed with the wholly isotactic chains
and depress the melting point only slightly. With the metallocene katalyst the
number of irregularities is much smaller, but they are randomly distributed over
all chains. They, therefore, hinder the crystallization process much more than with
the old process.
72 From polymers to plastics

– A phenomenon related to the above mentioned factors is, that the melting point
depends on the temperature at which the crystal has been formed. At a lower
temperature of crystallization crystallization proceeds more rapidly and less
regularly, while the crystal size is smaller; this results in a lower Tm.

– Camparison Tg and Tm
Since Tg and T m are both primarily governed by the same chain properties,
namely chain flexibility and chain interactions, some correlation between these
two quantities can be expected to be present. In Figure 4.5 Tg and Tm are plotted
against each other for a number of polymers. As a first approximation it appears
– for symmetric chains Tg ≈ 2 Tm ,
– for asymmetric chains Tg ≈ 3 Tm,
in which Tg and T m are, as a matter of fact, expressed in Kelvin. In reality the
values are spread round this relation by 10 to 30 degrees (see also Qu. 4.5).

Tg = 2 Tm
K Tg Tg = 1 Tm


100 asymmetric chains

symmetric chains

0 100 200 300 400 500 600 700 K


Figure 4.5. Relation between Tg and Tm.

4.3. The crystallization process

4.3.1. Nucleus formation and crystal growth
At a temperature below Tm the free enthalpy of the crystal is lower than that of the
liquid; hence the polymer will tend to crystallize. A crystal can, however, only be
formed from a nucleus. The process of crystallization can thus be split-up into two
different processes: nucleus formation and crystal growth.
Nucleus formation. Sometimes nuclei are already present, e.g. as contaminations,
additives, or as deliberately added accelerators for crystallization. Sometimes nuclei
Crystalline polymers 73

have to be formed by the polymer itself; statistical fluctuations in density and chain
conformations may form a first onset for the formation of a crystal. The first case is
denoted as heterogeneous nucleation, the second as homogeneous nucleation. In
both cases the nucleus size must, at a certain temperature, exceed a certain minimum
to be able to grow further into a crystal. This can be understood as follows:
Suppose that a spherical nucleus with radius r is present in the liquid. We consider
the difference in free enthalpy, ∆G, between this crystal and an equally large fluid
sphere. ∆G is composed of two components, namely A, the decrease of free enthalpy
upon crystallization, and B, the increase in free enthalpy as a result of the formation
of the interface. When ∆G v is the difference in free enthalpy per unit volume
between crystal and liquid, and γ is the interfacial energy, then:

∆G = – 3 πr3 ∆Gv + 4πr2γ = A + B

Figure 4.6 shows the dependence on r for both contributions; with small values of r
its square is predominant and ∆G increases with increasing r; the nucleus will stop
growing and (with homogeneous nucleation) it disappears. From a certain value of r,
the critical nucleus size, rk , ∆G decreases upon growth; the nucleus is then stable and
continues growing. The value of rk can be easily calculated: at rk:

d∆G 2γ
–4πr2 ∆Gv + 8πrγ = 0
dr = 0
rk = ∆G



Figure 4..6 Competition between interfacial and crystallization energy.

To make this equation more accessible, we write:

∆G = ∆H – Tm∆S = 0

At the melting point Tm is ∆G = 0 and ∆S = ∆H/Tm. At a lower temperature, T,

74 From polymers to plastics

∆H Tm – T
∆G = ∆H – T∆S = ∆H – T = ∆H· T
Tm m

We substitute this in the equation for rk :

2γ 2γ·Tm
rk = ∆G = ∆H(T – T)
v m

An important aspect of this equation is the factor (Tm – T) in the denominator; this is
the degree of undercooling. It appears that, just below T m , very large nuclei are
required; with further undercooling smaller nuclei become stable and thus able to
initiate the crystallization process. A similar reasoning can be followed for
heterogeneous nucleation; here also, with decreasing temperature, more and more
small particles become active as nuclei.
Generally speaking, it appears that the rate of nucleation is zero at the melting point,
and increases with decreasing temperature.
The growth rate of a crystal from a nucleus is also strongly dependent on the
temperature. With low undercooling, this rate is low, because the difference in free
enthalpy between liquid and crystal, which is the thermodynamic driving force for
crystallization, is still low. When the temperature decreases, this driving force
increases and so does the rate of crystal growth. There is, however, a counteracting
effect: the growth of a crystal goes accompanied by the movement of large chain
parts, and even of whole chains, with respect to each other. These movements require
the loosening of chain entanglements in the same way as with the flow of a molten
polymer. The chain mobility, as required for crystal growth, is thus related to the
melt viscosity, and is strongly dependent on temperature and chain length.
As a result of these two counteracting effects the rate of crystal growth shows a
maximum at some tens of degrees below Tm , and decreases to zero (at Tg ) with
further cooling.
From the rate of nucleus formation and the rate of crystal growth the rate of
crystallization results, which can be schematically represented as the product of both,
as shown in Figure 4.7. The rate of crystallization is zero at Tg and at T m, and shows
a maximum somewhere in between, a maximum which is lower when the chains are
The maximum rate of crystallization, vmax, strongly differs for different polymers; it
is, e.g. very large for PE, much smaller for PETP and near zero for PC (see also Qu.
4.10 – 4.15).
Crystalline polymers 75

v rate of

rate of nucleus
rate of

Tg Tm T

Figure 4.7. Temperature dependence of the rate of crystallization.

The degree of crystallinity, obtained after cooling down from the molten condition,
depends on the ratio between rate of cooling and vmax. With rapid cooling vmax may
be insufficient to lead to a significant crystallinity; the polymer then remains
amorphous and passes, at Tg , into the glassy state. It then follows the line ABCD
instead of ABEF, which holds for a crystallizing sample at a low rate of cooling
(Figure 4.8). When the thus obtained amorphous sample is slowly heated up starting
from D, it will be able to crystallize, and, thereafter it enters into the melting region
at G (DCGBA).


amorphous C

100% crystalline

Tg Tm T

Figure 4.8. V(T) relation for semi-crystalline polymers.

4.3.2. Isothermal crystallization

An analysis of the mechanism of crystallization can best be carried out at a constant
temperature. A sample is rapidly cooled down from the melt to a temperature below
Tm and kept constant at that level. The crystallization process can then be studied by
measuring the volume as a function of time in a dilatometer. The result is a curve as
shown in Figure 4.9. The volume approaches to an equilibrium value at which the
maximum possible crystallization is reached, and the rate of crystallization can, for
example, be expressed as 1/t0.5, in which t0.5 is the time in which half of the route is
76 From polymers to plastics

covered. By doing so at various temperature levels between Tg and T m, the curve for
the rate of crystallization, as indicated in Figure 4.7, can be established.



t 0,5 t

Figure 4.9. Volume change during crystallization.

Further analysis of results from isothermal experiments is based on a simple model.

Suppose that we are dealing with spherical crystallites which grow in three
dimensions. It can safely be assumed that the rate of growth is such that the radius of
the sphere increases proportionally with time:
r = v·t constant radial rate of growth.

The volume of the sphere is then

4 4
V = 3 π r 3 = 3 π v 3 t3

To represent the total crystalline volume Vc, V has to be multiplied by the number of
spheres. With a constant number of nuclei (heterogeneous nucleation) this number is
constant, so:

Vc (:) t3

With homogeneous nucleation, new nuclei are being formed continuously; the
number of nuclei and thus of growing spheres is proportional to time:

N = vt so Vc (:) t3 × t = t4

If the crystals do not grow in three, but in two dimensions, e.g. as disks, a similar
reasoning leads to:
Vc (:) t3 for homogeneous nucleation,
Vc (:) t2 for heterogeneous nucleation,

In the same way with one-dimensional growth (as needles):

Vc (:) t2 respectively Vc (:) t

In general we thus find:

Crystalline polymers 77

Vc = b · t a

Vc cannot be measured directly but only the total volume V; the relation between Vc
and V follows from:

V = Vcryst + Vamorphous = Vc + Va = Vc +

m – Vcρ c m ρc
= Vc + = – Vc [ – 1 ]
ρa ρc ρa

in which m = total mass, ma = mass of amorphous part etc, ρc and ρa the densities.

The volume decrease from the initial condition is ∆V = V0 – V , in which V0 = m/ρa,

since initially everything is amorphous.

m m ρc ρc ρc
∆V = – + Vc [ – 1] = Vc [ – 1] = [ – 1]·bta = cta
ρa ρa ρa ρa ρa

This is a first-order approximation, only valid for the first stage of crystallization
(otherwise ∆V would increase to infinity). As soon as growing crystals touch each
other, their growth is limited. A better approximation is given by the Avrami

∆V = (∆V)∞·(1 – exp(–cta ))

For small values of t this expression is equivalent to the first approximation, since
then exp(–cta ) ≈ 1 – cta.
To obtain some impression of the way in which ∆V depends on time for various
values of a, Figure 4.10 gives a schematic example for three cases: a = 1, 2 and 4,
while in all cases a value for t0,5 = 1000 min has been chosen.

∆V a=1

t min
0 500 1000 1500 2000 2500 3000

Figure 4.10. Change of volume with three different Avrami exponents.

78 From polymers to plastics

A better impression can be obtained by rewriting the Avrami equation:

∆V ∆V
1– a
(∆V)∞ = exp(–ct ) ln (1 – ) = –cta

ln[–ln (1 – )] = ln c + a ln t

From experimental data on ∆V as a function of time the proper way of plotting

provides the value of a as the slope of a straight line (see also Qu. 4.22).
Finally the effect of orientation on the rate of crystallization should be mentioned. In
a strained, oriented condition crystallization proceeds much more rapidly. By
orientation the ordering is increased; the chains approach the crystalline condition
more closely. As a matter of fact there is a relation to the increase in melting point,
mentioned in § 4.2; under strain the amount of undercooling ∆T is increased, but this
gives only a partial explanation of the increase in the rate of crystallization.

4.4. Crystalline structure

4.4.1. Crystalline fraction
Since polymers are never completely crystalline, it is of importance to know, how
big, in a certain case, the crystalline fraction is. The crystalline volume fraction, ϕ c,
can be calculated from the specific mass, ρ , if the densities of the amorphous
fraction, ρa, and of the crystal, ρ c, are known, since:

ρ – ρa
ρ = ϕc·ρ c + (1 – ϕc)·ρ a of ϕc =
ρc – ρa

It can be easily shown that the mass fraction, ψc, is given by:

ρ – ρa ρ c
ψc = ·
ρc – ρa ρ

ρ c can be determined from X-ray diffraction diagrams; these provide the dimensions
of the unit cell of the crystalline lattice, and thus the sum of the atom masses in such
a cell together with its volume (see Qu. 4.20). ρ a can be measured at a number of
temperatures above (or somewhat below) Tm ; extrapolation to room temperature
gives the required value (unless Tg is passed!). The table below gives, for a number
of polymers, some characteristic values, such as the range in which ϕc is mostly
found, and an average value for the density.
Crystalline polymers 79

ϕc (%) ρc ρa ρ
PA 6.6 35–45 1.22 1.07 1.14
POM 70–80 – – 1.41
PETP 30–40 1.455 1.335 1.38
PBTP 40–50 – – 1.30
PTFE 60–80 – – 2.1
PP 70–80 0.937 0.834 0.905
HDPE 70–85 1.00 0.855 0.95
LDPE 45–55 1.00 0.855 0.92

Another method is based on the intensity of X-ray diffraction as a function of the

diffraction angle θ; the ratio of the areas of the crystalline peaks and the amorphous
background gives ϕc (Figure 4.11).


Figure 4.11. X-ray diffraction of a semi-crystalline polymer.

Finally ϕc can be determined from the heat of melting of a sample via calorimetry, in
comparison with the melting heat of the pure crystal, known from literature.

4.4.2. Crystal morphology

An example of a polymer crystal is given in Figure 4.12, namely of PE, in which the
chains should be imagined to lie in zig-zag shape perpendicular to the plane of
drawing; the figure gives the projection of the C- and H-atoms. The crystal is
orthorhombic with a = 0.740 nm, b = 0.493 nm; the vertical dimension of the cell, c
= 0.253 nm. Also other polymers can crystallize with extended chains, such as PA-6
(monoclinic) and PA-6.6 (triclinic).

Isotactic vinyl polymers –(CH 2 –CHX)– can, because of sterical hindrance, in most
cases not lie in zig-zag shape; the chain then takes, as the most regular conformation,
a helical shape. An example is given in Figure 4.13, again as a projection on the
plane of drawing, for PP, which as a helix crystallizes in a monoclinical lattice.
80 From polymers to plastics

H H CH 3
C H c

Figure 4.12 Crystal lattice of PE. Figure 4.13. Projection of PP chain.

Because of the incomplete crystallization in polymers, crystalline and amorphous

regions are spatially arranged in some way. The oldest model for this arrangement is
of Hermann and Gerngross (1930), the so-called fringed micel structure (Figure
4.14). In this model the crystalline regions are embedded in an amorphous matrix,
while the chains pass from one region into another.
In an other model, the paracrystalline, (Hosemann, 1936), the amorphous regions are
embedded within the crystals, so the other way round (Figure 4.15). Both models are
realistic, and occur in combination with each other.
A detailed model for the single crystal was introduced by Keller in 1957 for
polyethylene. The chains (Figure 4.16) are more or less regularly folded in lamellae.

Figure 4.14. Fringed-micel model. Figure 4.15. Paracrystalline model.

Figure 4.16. Lamellae models.

The length of folding, so the thickness of the lamella, appears to be greater as the
temperature of crystallization has been higher. The most remarkable is, that the
lamellae increase in thickness upon recrystallization at a temperature nearer to Tm!
Crystallization from a stirred solution of, e.g. PE, results in a strongly oriented
bundle of fibres, on which, later on, transverse lamellae grow, the so-called shish-
Crystalline polymers 81

kebab structure (Figure 4.17). This observation (Pennings) forms the basis for the
recently by DSM developed super-strong PE-fibre.

Figure 4.17. ‘Shish-kebab’ model. Figure 4.18. Spherulite.

In polymers crystallized from the melt, in most cases spherulitic structures are
observed: spherical agglomerates of crystals and amorphous regions, grown from a
primary nucleus via successive secondary nucleation (Figure 4.18). The dimensions
of the spherulites are commonly between 5 µm and 1 mm. When spherulites grow
during the crystallization process, they touch each other and are separated by planes.
In a microtome slice they show a very attractive coloured appearance in polarized

4.5. Effect on properties

Crystallization brings about a drastic change in the pattern of properties. compared
with amorphous polymers. Upon the scheme: glass - rubber - liquid now the scheme
crystal - liquid is superimposed, since an amorphous phase remains always present.
This can be demonstrated easiest in the log E – T diagram. The amorphous polymer
follows curve a in Figure 4.19; two cases have been indicated: a1 and a2 , for a lower
and a higher molar mass, respectively.

log E


a1 a2

Tg Tm T

Figure 4.19. Effect of crystallization on E-modulus.

A fully crystalline polymer would follow curve b: no glass transition occurs and the
modulus drops sharply at Tm , either to zero or to a tail of the rubbery region (a 2).
82 From polymers to plastics

This picture is oversimplified; strictly speaking, curve b should be drawn at a

somewhat higher level than a in the glassy region; the stiffness of the crystal exceeds
that of the glass.
A partially crystalline polymer could follow curve c; at Tm the amorphous part passes
into a rubber, the crystalline part is unaffected. The actual curve resembles more
curve d, i.a. as a result of a broad melting region.
Some consequences of crystallization on the practical behaviour of polymers are:

– Above Tg the polymer can be used as a solid material to just below Tm, though its
stiffness is lower than that of a glassy polymer, and much more temperature
dependent (compare, e.g., E(PP) ≈ 1400 MPa and E(HDPE) ≈ 1000 MPa, with
E(PVC) ≈ 3200 MPa).

– The level of E between T g and T m strongly depends on the degree of

crystallization. This is clearly demonstrated by PE, where the density varies
between 0.91 and 0.97, the crystalline fraction between 0.40 and 0.85, and the E-
modulus between 100 and 1200 MPa at 20 °C.

– The rubbery region is wholly or partially masked by the crystallinity; the elastic
response of the melt is, therefore, much less pronounced.

– Parallel to the stronger temperature dependence of E for semi-crystalline

polymers is the stronger dependence on time; they show a higher tendency to
creep than amorphous, glassy polymers, (Figure 4.20), at least at temperatures
above or not too far below Tg .

% 23 °C POM (crystalline)
2 strain 20 MPa

PC (amorphous)

10 –1 1 10 10 2 10 3 10 4 10 5 hours

1 min 1 hour 1 day 1 year

Figure 4.20. Creep of an amorphous and a semi-crystalline polymer.

– A crystalline polymer is a two-phase system; for T < Tg it contains a glassy phase

and a crystalline phase, with, in general, a low impact strength. For T > T g the
amorphous phase is in the rubbery condition, which results in a much higher
impact strength.
Crystalline polymers 83

– Because of their to-phase nature, semi-crystalline polymers are, in most cases,

nontransparent; the difference in refraction index is responsible for strong light

– Crystallization of oriented chains is, in various respects, important for the

polymer properties. The fact has been mentioned before, that stereospecific
rubbers such as cis-1,4 polybutadiene can crystallize when under strain. The
spontaneously formed crystals contribute strongly to the strength of the
vulcanizate. A vulcanized natural rubber has, without carbon black reinforcement,
a tensile strength of about 40 MPa, whereas an unreinforced SBR breaks at about
3 MPa. (With SBR a high tensile strength can only be reached with carbon black.)

– Another effect of orientation and crystallization is found with fibres.

Crystallization of oriented chains brings about a stiffness and a strength which are
several times greater than those in the unoriented condition, though, as a matter of
fact, only in one direction.

– To conclude with, an overall picture is given in Figure 4.21 of various polymers,

ranged after Tg : the points for the crystalline polymers are connected by a vertical
line to their melting points Tm. In this survey the nature of the various polymers is
easily recognizable: For rubbers T g is far below ambient temperature To ; for
amorphous thermoplastics Tg lies amply above To . For semi-crystalline
thermoplastics Tg can be below as well as above To; in the latter case the
materials have an extra resistance against higher temperature.
84 From polymers to plastics




–100 BR


Figure 4.21. Survey of glass- and melting points.

4.6. Liquid-crystalline polymers

In § 4.3 we have already seen that polymers, in the rubber or fluid condition,
crystallize much more rapidly when their chains are oriented. Therefore a stretched
rubber, if stereospecific in its molecular structure, is able to crystallize at a
temperature considerably above its equilibrium thermodynamic melting point. Also a
thermoplast such as polyethylene, when in the molten state or in solution, can
crystallize spontaneously when the chains are being orientated in elongational flow.
The latter case is utilized when polyethylene is spun from a diluted solution (gel
spinning process), resulting in fibres of super-high strength and stiffness
(“Dyneema” fibres).
While with PE the chain is very flexible, a similar situation occurs with very stiff
chains; these have, as a result of their high characteristic chain stiffness (see § 2.4), a
much smaller number of conformation possibilities. Strongly exaggerated, they
Crystalline polymers 85

behave more as stiff rods than as coiled chains, in particular in elongational flow, in
which the “rods” are easily oriented and obtain a parallel direction. In such a flowing
liquid a considerable ordering is already present, somewhat like a crystal structure,
though this ordering only holds for the orientation and not for the position of the
chains, as in a real crystal. For such a situation the expression “liquid crystal” is
used. Upon solidification from the melt or the solution a material is formed which is,
in the direction of orientation, extremely strong and stiff.
Two types of liquid crystal polymers (LCP’s) can be distinguished: lyotropic and
thermotropic. The first type is formed in a solution; this is done when the polymer
has such a high melting point (low ∆S, see § 4.2), that it cannot be handled in the
molten condition without being degraded. If this limitation is not present, then the
orientation can be brought about in the melt; in such a case we have a thermotropic
A well-known example of a lyotropic LCP is the aramide fibre (Twaron or Kevlar),
an aromatic polyamide with the structure:

|| ||
–[–N– –N–C – –C –]–
The chain is exceptionally stiff because of the short-distance sequence of the phenyl
rings. The fibre is being spun and stretched from a solution in highly concentrated
sulphuric acid, and shows (in its length direction) nearly unsurpassed mechanical
An example of a thermotropic LCP is a copolymer of polyhydroxy benzoic acid and
polyethylene terephtalate:

|| || ||
–[–O– C – –]– and –[–C – –C –O– CH2 –CH2 –]–

The more flexible comonomer is introduced to reduce the melting point, so that the
polymer can be processed in the molten condition, e.g. in injection moulding and
extrusion. The end products show superior mechanical properties, but are, as a matter
of fact, highly anisotropic.

Rubbery and liquid phases

5.1. Rubbery phase

Above the glass-rubber transition temperature, Tg, large parts of the chain are free to
move; their thermal energy is high enough to overcome the interaction forces, and
the free volume increases with increasing temperature. The polymer is , however, not
yet in the liquid condition; the coiled chains are mutually entangled, in contrast to a
low- molecular amorphous material above its Tg (Figure 5.1):



Figure 5.1. A polymer is not a liquid just above its Tg.

Dependent of the amount of entanglement, thus of the chain length, a temperature

region exists in which the polymer does not flow, but is very soft, rubbery and
elastic. Though the chain entanglements are not permanent, because they are being
disrupted with increasing temperature and also with increased time of loading, they
act as temporary, physical cross-links.

We now understand that the material is, in this condition, soft and still coherent, but
the question is, for non-cross-linked polymers above Tg as well as for vulcanized
rubbers: why does it show spontaneous recovery after being stretched? As a model
we take a chain with freely rotating links; stretching of such a chain from a coiled
state does not require any energy, while, after stretching, it retains its shape.

This model, however, is only valid for chain molecules at zero Kelvin. When T ≠ 0
the chain parts possess thermal energy; vibration causes them to move in random
directions, which always results in a contraction of the stretched chain. The chain
tends to a state of higher probability, and eventually reaches a fully unoriented
random conformation, as described in § 2.4 (“random walk conformation”). To
Rubbery and liquid phases 87

prevent this, a force is needed, which is greater with stronger vibration and thus with
a higher temperature.

This implies the existence of a finite modulus of elasticity, and, also, of elastic
recovery. Since it concerns a competition between order and disorder, and has
nothing to do with any energy barrier, this is entropy elasticity. No change in energy
is present upon stretching in the rubbery state , but the entropy changes; in the
stretched condition the chains have much less conformation possibilities, and thus a
lower probability, W, and a lower entropy S = k·ln W (see also Figure 4.3). The
system tends to increase its entropy by retracting. From this reasoning we could now
try to calculate the E-modulus in the rubbery state.

Increasing the length dimension from l to l + dl requires a force K and an amount of

work dF = K·dl. This work equals the increase in free energy of the system, F = U –
T·S. Since no interactions are active, U does not change, so
dF dS
K = dl = –T

The next question is, how S varies with l. This can be calculated in a purely
statistical way. It is possible to count, for a schematized chain, the number of
conformations between two cross-links as a function of the distance between these
cross- links. Assuming that the increase of this distance is proportional to the total
strain, the calculation proceeds as follows:

Suppose the network is ideal, with N chain units between cross- links, while the
initial dimensions are: lx0 , ly0 and lz0 (Figure 5.2). The network is then deformed to
the dimensions l x , ly, lz. Statistical calculation of the number of possible chain
conformations between the cross-links results, with S = k·ln W, in:


ly0 ly

lx0 lx

Figure 5.2. Deformation of a rubber network.

∆S = –k·N·[(λ x 2 +λy2 +λz2 – 3)/2 – ln λxλ y λ z]

in which λx = lx/lx 0 etc., the deformations in the three main directions, with εx = λx –
1 etc. When the strain is unidirectional in the x-direction, then
88 From polymers to plastics

λ y = λz ≠ λx,


λ y 2 = λz2 = 1/λx ,.

Since the volume

V = λ x λ y λ z·V0,

is, for rubbers, practically independent of the strain (the Poisson ratio ν ≈ 0.5), is

λxλyλz = 1 ,


λx2 1 3
∆S = –k·N·[ 2 + – 2]

The force is
dS d∆S
K = –T dl or K = –T dl

or, with l = lx 0 ·λ :
T d∆S NkT
K=– =
l x 0 d λ x lx 0
(λx – λ1 2 )

The stress σ is:

σ = N*kT·(λx2 – )

in which N* is the number of chain units per unit volume.

For small strains: λx = 1 + εx, λ x 2 ≈ 1 + 2ε x , 1/λx ≈ 1 – εx, so that:

σ = N*kT·(1 + 2εx – 1 + εx) = 3N*kT·εx


E = σ/ εx = 3N*kT

For large strains the expression for the stress can be simplified into:

σ = N*kT(λ x )2
Rubbery and liquid phases 89

In most cases the stress is not expressed as force per unit of the actual area of cross-
section, but is, rather, based on the original area; in that case it can easily be shown
that then the stress is given by:

σ = N*kT·λ x

This relation holds for strains up to 300 or 500 % (λ = 4 to 6); with even higher
strains some chain parts may approach their fully stretched conformations; moreover
crystallization may occur. In both cases the slope of the stress-strain curve increases.
In Figure 5.3 the resulting diagram is schematically presented.



0 1 2 3 4 5 6 7 8 λ
0 1 2 3 4 5 6 7 ε

Figure 5.3. Stress - strain diagram of a rubber.

The initial modulus of elasticity, E = 3N*kT, can also be written as

3ρ RT

where ρ is the density, R the gas constant and Mc the molar mass between cross-
links, since:
k N*·Mc
NA = ρ
N*Mc· =

(NA = Avogadro-number)

The initial E-modulus, and also the stresses at higher strains, appear to be
proportional to the absolute temperature, as our simple model of the vibrating chain
already suggested. For an ideal vulcanized rubber, showing only entropy elasticity,
this can, indeed, be observed. A simple experiment on a thin tube, stretched by a
dead weight to several times its initial length, through which, alternatively, water of
100 °C and of 0 °C flows, clearly shows the “negative coefficient of expansion” of
strained rubber, though the theoretical ratio of the strains, 373/273, is not achieved.

Deviations from the proportionality σ (:) T result from, a.o. three complications,
which have been neglected in the simplified treatment:
90 From polymers to plastics

K 400 K 400
% %
300 200
10 5
Tg T Tg T

Figure 5.4. Deviations from ideal rubberelastic behaviour.

– First, some energy elasticity is always present; the lines in a plot of K against T
(Figure 5.4), when extrapolated back from the glass transition temperature, do not
pass through the origin; the intercept with the K-axis represents the small effect of

– Secondly, the coefficient of thermal expansion of the unstrained rubber has not be
taken into account; its effect is evident from a decrease of the slope of the K-T
curve; below a certain strain (the “thermodynamic inversion point”) this slope is
even negative.

– In the third place, we have, for simplification, considered the free energy F = U –
TS instead of the free enthalpy G = F + p V; the volume, though approximately
constant upon straining, still shows a minor increase.

It should also be remarked that we have supposed the existence of permanent cross-
links; this may hold for vulcanized rubbers, but in unvulcanized condition the
entanglements are gradually loosened under strain, to be formed again in other
places. As a consequence, an increase of E with T is, for uncrosslinked polymers,
hardly observable; in most cases E decreases with increasing temperature, which
goes together with a gradual transition from the rubbery into the liquid state.

The E-modulus appeared to be inversely proportional to the distance between cross-

links, so, approximately, proportional to the number of cross-links. If no permanent
cross-links are present, i.e. in the unvulcanized condition, then E is governed by the
number of physical entanglements. Light vulcanization does not contribute
significantly to total number of cross-links; therefore, in lightly crosslinked
vulcanizates, such as technical rubbers, the E modulus is hardly dependent on the
degree of vulcanization (Qu. 5.4 en 5.5).

Strongly cross-linked systems such as thermosets, however, show in the rubbery

phase a much higher modulus than a rubber, but still considerably lower than in the
glassy state. Figure 5.5 schematically presents how E changes with T for systems
Rubbery and liquid phases 91

log E

increasing crosslinking

Figure 5.5. E(T) for networks.

log E

increasing M

low molecular glass T

Figure 5.6. Effect of chain length on rubbery region.

with increasing degree of cross-linking; also the increase of Tg , discussed in § 3.4.1,

is visible in this diagram.

5.2. Transition of rubber to fluid

With increasing temperature the rubbery phase gradually passes into a fluid state; the
temperature region in which this occurs, is strongly dependent on the chain length
(see Figure 5.6). Longer chains have a greater number of entanglements and require
a higher temperature to become disentangled. Since the longest chains play a

predominant role, higher averages such as Mz govern the rubber-fluid transition.

Moreover, the time scale is of importance, since, at a certain temperature, an

entanglement can be loosened under stress after a somewhat longer time. Figure 5.7
presents, schematically, how log E varies with temperature for various loading time
scales; in this figure the dependence of Tg on the time scale is indicated, and, in
particular, the effect of time on the rubber- to-flow transition. On a short time scale
the polymer is, at temperature T 1 , strongly rubberelastic; at a long time sc ale it
behaves as a fluid. This dualism visco-elastic behaviour, has a large number of
practical technological consequences, some of which will be discussed in § 5.4.
92 From polymers to plastics

log E


1 10 –4 sec

Figure 5.7. Effect of time scale on the rubber - liquid transition.

5.3. Fluid condition

5.3.1. Viscosity
In the fluid condition the chains move as a whole with respect to each other. The
many entanglements have to be disrupted continuously; therefore, the viscosity of
liquid polymers is very high. In the table below a comparison is given of the
viscosities of various liquids, expressed in Pa·s (= N·s/m2 = 10 poise) at 20 °C unless
stated otherwise.

water 0.001
alcohol 0.0012
fluid metals 0.001 – 0.002
glycol 0.02
linseed oil 0.05
olive oil 0.08
machine oil 0.1 - 0.5
glycerol 1.5
polymers 100 – 10.000
(at processing temperature)

The viscosity of polymers is strongly dependent on their molar mass, namely

according to η (:) Mw3,4 (weight average). The high value of the exponent, 3.4,
reflects the large influence of chain entanglements on the flow behaviour; when the
chains are twice as long, the number of entanglements has increased so much that the
melt viscosity is more than ten times higher!

As a matter of fact, the viscosity is a very dominant property for the processing
technology. A fluid polymer mass with a high viscosity requires high pressures to be
transported in processing machines. How strongly the processability prevails, has
already been discussed in § 2.2.3; for some thermoplasts like PE, PP and ABS,
Rubbery and liquid phases 93

grades differing in molar mass. are usually characterized by their melt index, which
is the amount of material flowing through a standard capillary, as a first measure for
their processability.

With rubbers a similar situation is met, but now with the aid of a rotation viscometer.
The Mooney viscosity is measured as the torque needed to rotate two parallel plates,
between which the rubber mass is present, with respect to each other. This provides a
rough indication of the viscosity, and thus of the molar mass. This measurement can
also be used to characterize the vulcanization behaviour; under vulcanization
conditions the increase of the Mooney viscosity indicates the onset of network
formation. When the network develops further, a continuous rotation can, of course,
no longer be applied because the viscosity increases unlimitedly; therefore an
oscillation method is mostly used with a cone-and-plate geometry. Initially, the
viscosity is being measured, and later on the build-up of the E-modulus of the
network. Another characterization of the viscosity of unvulcanized rubbers is the
Hoekstra method. The rubber is present between two parallel plates, which are
moved towards each other with a certain speed; the force needed to do so is an
indication of the viscosity.

These simple standard methods for characterizing the viscosity only provide a first
impression of the processing behaviour. In reality considerable complications occur.
The first is that we are already in trouble when we try to define the viscosity
properly, as a result of th e non-Newtonian behaviour.

5.3.2. Non-Newtonian behaviour

surface force K
area A speed v
h γ

Figure 5.8. Definition of viscosity.

When using the word “viscosity” it is implicitly assumed that stress and shear rate
are proportional. This means that the force K required to move two plates, between
which a liquid is present, parallel to each other with a speed v (Figure 5.8), is
proportional to v.

To eliminate the effect of the dimensions, the force is divided by the area of the
plates, A, so as to give the shear stress:

τ = K/A

and the speed by the plate distance, h, resulting in the velocity gradient v/h. The
94 From polymers to plastics

latter equals the rate of shear γ/t or dγ/dt or γ. The proportionality thus reads:

K (:) v and τ = K/A (:) v/h = γ

The proportionality constant is defined as the viscosity η:

η = τ/γ

When τ and γ are, indeed, proportional to each other, the liquid is called Newtonian.
However, various fluids deviate from Newtonian behaviour, as indicated in Figure

b a
τ c

a = Newtonian
d b = plastic (yield stress)
c = pseudoplastic
d = dilatant

Figure 5.9. Possible deviations from Newtonian behaviour.

Fluid polymers mostly show pseudoplastic behaviour: the rate of flow, γ, increases
more than proportionally with increasing shear stress τ. A unique viscosity value can,
therefore, not be defined; when, for instance, with a certain value of γ, a value of τ is
measured, then

η = τ/γ is nothing more than an apparent viscosity

only holding for a single level of τ or of γ.

Full characterization of the flow behaviour, therefore, requires the determination of

the apparent viscosity over a broad range of τ or of γ. For low rates of shear η can be
measured with a cylindrical or a cone-plate rotation viscometer. For higher rates of
shear (from about 1 sec –1) capillary viscometers can be applied.

The apparent viscosity may, in the region of shear rates occurring in practice,
decrease by a factor of 10 to 500 with increasing γ. In most cases η is represented as
a function of τ (or, also, of γ) on a log-log scale. An example is given in Figure 5.10
for three grades of the same polymer.

A en B have the same viscosity at low stress: the “zero-shear viscosity”. They have,
— —
therefore, the same Mw, since the relation given before: η = Mw3.4, is valid for this

zero- shear viscosity. C has a ten times lower value; its Mw, is thus about twice as
Rubbery and liquid phases 95

10 10 –1 1 10
(Pas) γ 10
10 (s–1)

A B 10
10 4

2 3 4 5 6
10 10 10 10 10
τ (Pa)

Figure 5.10. Non-Newtonian behaviour of three grades of the same polymer, differing as to
molar mass and -distribution.

The difference between A and B is due to a difference in molar mass distribution

MMD; for A this is broader than for B (and for C). A broader MMD causes the
polymer to deviate more strongly from Newtonian behaviour. A first approximating
— —
measure of the width of the distribution is Mw/Mn ; as far as the effect on the non-
— —
Newtonian behaviour is concerned, Mz/Mw would be a better measure.

It further appears that, with higher rates of stress, the effect of the MMD may

overrule the effect of Mw on the apparent viscosity: the curves for A and C intersect.
From Figure 5.10 one might conclude: there is a “Newtonian region” for A up to γ =
0.1 sec–1 or to τ = 103 Pa, and for B up to γ = 1 sec–1 or to τ = 104 Pa. In this
“Newtonian region” the viscosity would be “constant”. In how far this is indeed true,
is apparent from the linear plot in Figure 5.11. As happens more often, the use of a
log-scale may lead to a distorted look on reality.

η Pas
10 000


100 000 τ 200 000 Pa

Figure 5.11. Where is the “Newtonian region”?

To represent the non-Newtonian behaviour of polymers in a formula, the “power-

law” is often used
96 From polymers to plastics

τ = k·( γ )n

The exponent, n, the “power-law-index” is n = 1 for Newtonian behaviour, and n < 1

for pseudoplasticity. The apparent viscosity is then given by:

η = τ / γ = k·( γ )n – 1
or by
η = k1/n·τ(n – 1)/n

These expressions provide a reasonable approximation in the region where the

curves, as e.g. in Figure 5.10, do not deviate too much from a straight line, so with A
above 1 sec–1 or τ > 3000 Pa. For lower values the power-law fails, in particular
because the viscosity for τ = 0 or γ = 0 would become infinitely high. Various other
approximations have been proposed to meet this objection; when calculating the flow
in a processing machine, one chooses the formula which gives the best fit in the
relevant range of stresses.

The temperature dependence of viscosity is often expressed by an Arrhenius


η = A·exp(E/RT)

In practice, most polymers vary by a factor between 1.5 and 5 over a temperature
difference of 30 °C. With non-Newtonian behaviour it is important how the
temperature dependence, e.g. the value of the energy of activation E, is defined,
namely for constant shear stress or for constant shear rate. Constant stress τ gives, in
general, a more constant value for E, since the log η – log τ curves shifts, upon
temperature change, in the vertical direction (just as with a change in M), whereas,
with constant γ the pattern is more deformed (see Figure 5.12).

log η
T1 < T2 < T3

T2 T1


n st
t log τ
τ constant

Figure 5.12. Temperature dependence of viscosity.

The deviation from Newtonian behaviour described above can also be related to the
melt elasticity (see § 5.3.3), since this is a rubber-elasticity, which results from an
Rubbery and liquid phases 97

entropy decrease when the most probable, chaotic shape of the chain is transformed
by orientation into a sigar-shaped one. These chains slide more easily along each
other in a shear field because their mutual entanglements are decreased, which
results in a lower viscosity.

When does this process start to play a significant role? When the relaxation time of
such an elastic deformation exceeds the time scale at which the deformation takes
place, which is the reciprocal shear rate 1/ γ . We have seen before that for a number
of similar polymers with the same shape of molecular mass distribution, the
deviation from Newtonian behaviour starts with the same value of the shear stress,
. .
thus (according to γ = τ /η ) at values of γ which are inversely proportional to the
(zero)viscosity η . It seems plausible to suppose that (again with similar mol mass
distributions) the relaxation time of the elastic deformation is proportional to the
viscosity (see also next section), so that the above mentioned observation is
— — —
With a broader mol mass distribution, Mz en Mz+1 will increase at the same Mw,, and,
therefore, the elastic response will become more pronounced. This means: a longer
relaxation time and a lower value ofγ (and of τ) at which the apparent viscosity
decreases. The log η - log τ curve is then shifted to the left.

In fact, we have to speak of a spectrum of relaxation tmes (see, e.g. Figure 6.8), in
which the value of γ at which the first deviation from Newtonian behaviour appears,
is a (reciprocal) measure of the longest relaxation time.

5.3.3. Melt elasticity

As discussed before (§ 5.2), a molten polymer shows also elastic behaviour,
particularly on a short time-scale: the fluid is visco-elastic. This can, in a simple
experiment, be demonstrated in two ways. When we let a bar rotate around its axis in
a viscoelastic fluid, then, after removal of the driving torque, it will rotate back over
a certain angle. Moreover the fluid will, during rotation, creep upward along the bar,
which indicates the existence of normal stresses next to shear stresses.

The elastic behaviour is more strongly pronounced with higher molar mass; higher
— —
averages, as Mz and Mz+1 , play the most important role; the longest chains more or
less tie the other chains together.

5.4. Some consequences for processing

Most processing techniques are carried out in the fluid phase. Here the melt viscosity
is the most important parameter. Often this is expressed as the melt index (m.i., see
§§ 2.2.3 and 5.3.1), which is a measure for the reciprocal viscosity for a given value
98 From polymers to plastics

of shear stress. A higher m.i., so a lower viscosity, indicates a better processability.

This is, e.g. evident for injection moulding, where for longer flow paths grades with
a higher melt- index are being recommended, and for film extrusion, where a higher
melt index enables the production of thinner films. An example of the latter case is
given in Figure 5.13, derived from a list of PE-grades supplied by DSM, in which the
minimum attainable film thickness is plotted against the melt-index. The grades
indicated as 1 and 3 do not fit into the pattern; this will be discussed later on
100 d min

50 2

30 3 5

melt index 8
0,2 0,3 0,5 1 2 3 5

Figure 5.13. Minimum film thickness as a function of melt-index.

The melt-index is merely a one-point determination of the flow behaviour and is not
a sufficient criterion in comparing grades with different molar mass distribution.
This is apparent from Figure 5.10 and also from Figure 5.14, in which two grades of
the same polymer with the same melt index but with different MMD are
schematically represented. The curves intersect at the shear stress which is is present

in the melt-index measuring device (2·104 Pa). A has a higher Mw but a broader
distribution. For injection moulding A is more attractive, since the mould filling
process in injection moulding takes place at very high rates of shear. The more
strongly pronounced elastic behaviour, however, increases the risc of frozen-in
(rubberelastic) deformations, resulting in anisotropy, and, therefore, in poor stability
of shape and dimensions, and in brittleness at the point of injection. The latter is due
to the fact that the strength perpendicular to the orientation direction is lower.

Type B in Figure 5.14 is typically suited for rotational moulding. The stresses and
the shear rates in this process are very low: after heating the particles flow together
under the influence of their own weight and surface tension only, so that the zero-
shear viscosity should be as low as possible. For a given melt index this means a
narrow MMD. This also implies that, for a given M w (which governs the viscosity),

Mn is higher. Properties which are highly desirable for rotation moulded articles,
such as a good impact strength and a high resistance against environmental stress
Rubbery and liquid phases 99

cracking, largely depend on the number of chain ends and thus on Mn .

log η A
melt index

log τ

Figure 5.14. Two grades with the same melt index.

Transportation of a molten polymer through a converging channel, e.g. with

extrusion, goes accompanied with elongational flow; when the rate of elongation is
high, next to the viscous deformation also a considerable elastic component may be
present. After the polymer leaves the channel, the elastic deformation will
spontaneously recover, which is apparent as an increase in diameter of the extrudate,
the so-called die-swell.

The rubber-elastic deformation of the liquid stream can reach such high values that
the liquid breaks. Though the remaining fragments will stick together in a further
part of the die channel, this “melt fracture” forms a notorious limitation in the rate of
production with extrusion processes. The only remedy is to increase the time scale of
deformation, so that the elastic behaviour is less outspoken.

Rapid production in e.g. injection moulding, requires a high rate of cooling from the
melt; when the melt is still deformed in a rubberelastic way, these deformations are
frozen-in as chain orientations and they give rise to anisotropy in the properties of
the end product. On the other hand, the freezing-in of orientations is utilized in the
production of fibres and oriented films; freezing-in is, in these cases, achieved by
crystallization. Besides the improvement in strength and stiffness, brought about by
these orientations, they may be applied in shrinkage films for packaging purposes.
The lower the melt index, the better the polymer is suitable for this purpose. One of
the deviating points (3) in Figure 5.13, denoting a grade with an extra low minimum
film thickness, belongs to a grade with an extra broad MMD; its flow is better than
would be expected from its melt-index and its elastic behaviour is more strongly
pronounced; it is not surprising that this grade is specially recommended for
shrinkage film. The other deviating point in Figure 5.13 (1) probably has an extra
narrow MMD, as estimated from its poorer flow for a given melt index. This grade is
recommended for its very high tear strength and toughness, which are, in particular,

controlled by Mn . Here again, the information provided can be joined together into a
simple pattern!

Another example concerns polycarbonate PC. Here also grades of several levels of M
100 From polymers to plastics

are available. Of actual importance is the type with the lowest M, with the lowest
melt viscosity, for the manufacture of compact discs, where extremely fine
reproduction of details is required. The low M is also responsible for the lowest
possible frozen-in (rubberelastic) deformations; this is essential to reduce the optical
anisotropy, the birefringence, which is catastrophic for the quality. The impact
strength is certainly not promoted by the choice of a low M grade, but PC has such
an exceptionally high impact strength that this does not form an objection.

A clear example of the effect of orientations during flow is the injection moulding
behaviour of rubbers with a sterospecific chain structure such as NR and IR. These
rubbers can, when under strain, crystallize spontaneously (which is very useful for
the properties of the vulcanizate, see § 4.5). In a processing operation, such as
injection moulding, this phenomenon may, however, play a negative role;
orientation-induced crystallization in the converging channel may cause an
intolerable increase of viscosity (Figure 5.15). With natural rubber, NR, (100% cis)
injection-moulding is, therefore, much more difficult than with a synthetic isoprene
rubber (IR) with, e.g., 92% cis content. In practice, for such an IR only about 40 %
of the injection pressure needed for NR suffices.

10 7 Cis 1-4 polyisoprene

(Pa·s) 100% cis
96% cis
10 6 NR 50 °C

10 5 IR
92% cis

10 4
10 2
1 10 γ 10 sec–1

Figure 5.15. The flow behaviour is affected by crystallization.

Some processing operations are mainly carried out in the rubbery phase. With sheet
forming, such as vacuum forming, the sheet should possess rubbery properties to be
manageable on the machine. Consequently, the deformation is partly or wholly
rubberelastic, which can be easily demonstrated by complete recovery to the initial
shape after heating to above Tg (or Tm). With calendering we have a similar situation:
for taking-off the sheet from the last roll a pronounced rubberelastic behaviour is
essential. The processability with these techniques, therefore, depends on the length
of the rubbery region on the temperature scale. Amorphous polymers do not offer
any problem; for crystalline polymers the rubbery region is partly or wholly masked
by the crystallinity (see § 4.5). Only with high molar masses may such polymers,
Rubbery and liquid phases 101

such as PE and PP, be vacuum formed under strictly controlled temperature

conditions; for polyamides and polyesters, however, the chains are not long enough.


6.1. Models
The expression “visco-elastic” signifies the dual nature of a material: on the one
hand it behaves in a viscous way, as a liquid, on the other hand elastically, as a solid.

For an ideal solid, Hooke’s law holds: the stress, σ , applied is proportional to the
deformation, ε, and the proportionality constant is the modulus of elasticity E, so σ =
E·ε . Besides E also other quantities play a role, such as the shear modulus, G, in a
shearing deformation or torsion, which is related to E. For the sake of simplicity we
shall mainly use E as a representative quantity for the elastic stiffness in any
geometry of loading.

As a model for E we take a helical spring with stiffness E. The response ε of such a
spring to a stress σ is schematically indicated in Figure 6.1; the response is
instantaneous, without any time dependency, and the recovery after release of the
stress is also instantaneous and complete.

σ ε


t t

Figure 6.1. Response of an ideal spring.


arctan σ
η t

Figure 6.2. Response of an ideal liquid.

For an ideal liquid Newton’s law holds: the stress is proportional to the rate of
Visco-elasticity 103

deformation dε/dt = ε ; the proportionality constant is the viscosity η , so σ = η·ε. As

a model we choose a dashpot; within a cylinder filled with a fluid a piston can move
with some clearance. Figure 6.2 shows its behaviour: There is no instantaneous
response; the deformation is proportional to time, and no recovery takes place. The
dashpot is characterized by η .

For viscoelastic materials combinations of these two models can be used, e.g. a
spring and a dashpot in series or parallel. The first combination is called the Maxwell
element; its response under constant stress is the sum of that of its two components:

σ ε

ε2 ε1
ε1 ε2
t t

Figure 6.3. Response of a Maxwell element.

σ σ 1 t
ε = ε1 + ε2 = E + η ·t = σ [E + η]

i.e. a spontaneous elastic deformation which returns when σ = 0, plus permanent

flow. This element might, very roughly, represent the behaviour of a molten
polymer, since we have seen that, besides the normal fluid behaviour, also an elastic
component is present, which recovers upon unloading.

Next to this case of creep under constant loading, we also consider the stress
relaxation which occurs when the deformation is kept constant. At t = 0 the model is
jumpwise deformed to a strain ε. The instantaneous response is a stress σ0 = E·ε ; the
spring is strained, the dashpot does not yet respond. The dashpot is, at t = 0,
subjected to the same stress, so it starts flowing, while it takes over an increasing part
of the imposed strain so that the strain in the spring, and also the stress, decrease.

At time t the deformation of the spring is ε1 en that of the dashpot ε2, while the stress
in both σ1 = σ2 . Now

dε2 d(ε – ε 1 ) d ε1
σ1 = E·ε1 = σ2 = η dt = η dt
= –η dt
ε 1 + ε2 = ε
104 From polymers to plastics

d ε1
–η = E·ε1

dε1 E
=– dt
ε1 η

ln ε 1 = – ·t + c

ε1 = exp(–(E/η)·t) · c ′

σ = E·ε 1 = E·exp(–(E/η)·t) · c′

At t = 0 is σ = E·ε , so σ(t) = E·ε·exp(–(E/η)·t) = E·ε·exp(–t/τ) in which τ = η/E , the

relaxation time. The relaxation time, as indicated in Figure 6.4, is the time in which
the stress is reduced to 1/e times its original value (about 37%). (See also Qu. 6.2 and


σ0 = E·ε
η τ=η
e t
t = –0 t = +0 t = t t=∞ τ

Figure 6.4. Stress relaxation of a Maxwell element.

A parallel array of E and h gives a Kelvin-Voigt element. This model does not allow
an instantaneous deformation (the stress on the dashpot would be infinite), and it
does not show stress relaxation. At a constant stress it exhibits creep; at time t its
strain is ε(t); the stress in the spring then is:

σ1 = E·ε(t)

in the dashpot

σ2 = η· dt

and the total stress, σ1 + σ2 = σ is constant.

Visco-elasticity 105

σ = E·ε + η dt

η dt = –E·ε + σ

The solution of this differential equation is:

ε(t) = E [1 – exp(–t/τ)]

with τ = η/E.

At t = 0, ε = 0; at t = ∞, ε = σ/E; this final deformation is approached asymptotically.

After taking away the stress at a new t = 0, while ε = ε1, it follows from:

σ = 0 = E·ε + η dt


η dt = –E·ε:

ε = ε1 exp(–t/τ)

in other words: the recovery proceeds asymptotically to ε = 0 with the same

relaxation time τ = η/E; see Figure 6.5. (see also Qu. 6.5).

σ ε

E σ/E

τ t

Figure 6.5. Creep of a Kelvin-Voigt element

Both models, the Maxwell element and the Kelvin-Voigt element, are limited in their
representation of the actual viscoelastic behaviour; the former is able to describe
stress relaxation, but only irreversible flow; the latter can represent creep, but
without instantaneous deformation, and it cannot account for stress relaxation. A
combination of both elements, the Burgers model, offers more possibilities. It is well
suited for a qualitative description of creep. We can think it as composed of a spring
E1, in series with a Kelvin-Voigt element with E2 and η 2 , and with a dashpot, η 1
106 From polymers to plastics


E2 η2 ε

ε tot

Figure 6.6. Creep of the Burgers model.

(Figure 6.6).

Under a constant load σ we can distinguish between three types of deformation:

– the spring: ε1 = σ/E1

– the Kelvin-Voigt element: ε2 = σ/E2 ·[1 – exp(–E2 ·t/η2)]
– the dashpot: ε 3 = σ·(t/η1) .

The total deformation is:

1 1 t
ε = σ[E + E (1 – exp (–t/τ) + ]
1 2 η1


which means

-– spontaneous elastic deformation,

– delayed elastic deformation, or reversible creep,
– irreversible creep (flow) (see also Qu. 6.7 en 6.8).

The models described so far provide a qualitative illustration of the viscoelastic

behaviour of polymers. In that respect the Maxwell element is the most suited to
represent fluid polymers: the permanent flow predominates on the longer term, while
the short-term response is elastic. The Kelvin-Voigt element, with an added spring
and, if necessary, a dashpot, is better suited to describe the nature of a solid polymer.
With later analysis of the creep of polymers, we shall, therefore, meet the Kelvin-
Voigt model again; in more detailed descriptions of the fluid state the Maxwell
model is being used.
Visco-elasticity 107

In a quantitative way, these simple models are not powerful enough to account for
the behaviour of a real polymer. One of their prime shortcomings is, that they
describe processes with a single relaxation time only. To illustrate this, we consider
the stress relaxation of a Maxwell model. In a σ -log t plot (Figure 6.7) the stress
relaxation curves for three different values of the relaxation time have been
indicated; with this way of plotting the three curves have the same shape and are
shifted with respect to the log t axis. Also a more realistic relaxation curve is drawn,
which extends over a much broader region of the log t scale than a single element.

σ E1 E2 E3 Ei

η1 η3 … ηi

τ1 τ2 τ3
log t τ τ2 τ3 … τi

Figure 6.7. Relaxation of three single Figure 6.8 Generalized Maxwell model.
Maxwell elements compared with actual

The reality is better approximated by a “generalized Maxwell model” (Figure 6.8),

consisting of a large number of Maxwell elements in parallel, each with its own
relaxation time, τi , and its own contribution, Ei , to the total stiffness. This system can
be described by:

σ(t) = ε·E1·exp(–t/τ1) + ε·E2·exp(–t/τ2) + … + ε·Ei·exp(–t/τi)

σ(t) = ε·∑ i Ei·exp(–t/τi)

A relaxation function is now defined as

E(t)= ∑ i Ei·exp(–t/τi)

We prefer to use a continuous spectrum H(t), the relaxation spectrum:

E(t) = ∫ H(t)·exp(–t/τ)·dτ
108 From polymers to plastics

in which H(τ)dτ is the contribution to the stiffness in the interval between τ and
τ + dτ.

In a similar way the creep can be represented by a number of Kelvin-Voigt elements

in series:

σ σ
ε(t) = E ·[1 – exp(–t/τ1)] + … + E ·[1 – exp(–t/τi)] + …
1 i
ε(t) n
= ∑ i [ ·(1 – exp(–t/τi))]
σ 1

To express creep, use of the reciprocal modulus, the compliance, D = 1/E, is to be

preferred; the creep function then reads:

ε(t) n
D(t) = = ∑ i Di ·[1 – exp(–t/τi)]
σ 1

or, as a continuous spectrum:

D(t) = ∫ L(τ)·[1 – exp(–t/τ)]·dτ

The spectrum L(τ ) is called the retardation spectrum. With the integration from τ = 0
to τ = ∞ the immediate elastic response and the irreversible flow are, respectively,
automatically taken into account.

With all these models, the simple ones as well as the spectra, it has to be supposed
that stress and strain are, at any time, proportional, so that the relaxation function
E(t) and the creep function D(t) are independent of the levels of deformation and
stress, respectively. When this is the case, we have linear viscoelastic behaviour.
Then the so-called superposition principle holds, as formulated by Boltzmann. This
describes the effect of changes in external conditions of a viscoelastic system at
different points in time. Such a change may be the application of a stress or also an
imposed deformation.

The superposition principle states that each change gives its own contribution to a
(later) effect, independent of the other changes. This is important, since not only the
magnitude of the change, but also the time elapsed to the observation of its effect,
influence the magnitude of the effect.

A simple example of the superposition principle is given in Figure 6.9. At t = t1 a

stress σ1 is applied, which is increased to σ2 at t = t 2 . The response of the system is
Visco-elasticity 109

such that the strain from t = t2 is a superposition of the strain as a result of a

continuation of σ1 and the one resulting from a stress σ2 – σ1, applied at t = t2.

Expressed in a formula:

ε(t) = σ1 ·D(t – t1) + (σ2 – σ1)·D(t – t2)

σ2 total
result of σ1
result of σ2 – σ1

t1 t2 t t1 t2 t

Figure 6.9. Superposition principle.

With an arbitrary, continuously changing stress σ(t), this can be written as:
t dσ
ε(t) = ∫ D(t – t′) ·dt′

For each small variation in stress (dσ/dt′)·dt′ at t = t′, its contribution to the
deformation is determined by the pertinent compliance D(t – t′).

Later on we shall see that the superposition principle is is, for polymers, only
seldomly obeyed; linear viscoelasticity is only met at very small stresses and
deformations; at loading levels occurring in practice the behaviour may strongly
deviate from linearity. However, the superposition principle provides a useful first-
order approximation.

The functions E(t) and D(t) are of importance for two reasons. First, they provide a
first impression of the response of a polymer under stress or strain as a function of
time; they are, therefore, of interest when plastics are used in load-bearing

On the other hand knowledge of these functions and of the spectra of relaxation (or
retardation) times derived from them, is very helpful for obtaining insight into the
molecular mechanisms by which they are originated. Analysis of the time
dependency of mechanical properties thus provides a powerful tool to investigate the
relations between structure and properties.

Creep and relaxation experiments are carried out on time scales ranging from several
110 From polymers to plastics

seconds up to several years. When knowledge of the viscoelastic behaviour on

shorter time scales is also required, then vibration experiments (dynamic mechanical
experiments) have to be carried out.

6.2. Dynamic mechanial behaviour

When a periodic deformation, ε = ε 0 ·sin ω t , is imposed upon a purely elastic
material, the stress will also be periodic according to σ = σ 0 ·sin ω t (the use of a
cosine function gives the same results). With viscoelastic materials, however, a
phase shift δ occurs:

σ = σ0·sin(ω t + δ)

This can be understood as follows: For an ideal spring δ = 0; for a pure liquid:
σ = η·(dε /dt) = η·ε0·ω·cos ωt = σ0·sin(ωt + 2 )
then δ =2 (90°); the stress proceeds in advance of the strain over a quarter of a
period. For a viscoelastic material δ is in between these extremes: 0 < δ < 2 . The
equation for σ can be worked out further:

σ = σ0·sin(ωt + δ) = σ0·[sin ωt·cos δ + cos ωt·sin δ] =

= (σ0·cos δ)·sin ω t + (σ0·sin δ)·cos ω t

This equation shows the existence of two components of the stress: the first one is in-
phase with the deformation, with an amplitude σ0·cos δ, the second one has a phase
difference of 90°, with an amplitude σ0·sin δ .

We can also define two moduli of elasticity σ/ε:

E1 (in-phase) = (σ0/ε 0 )·cos δ

E2 (90° out-phase) = (σ0/ε 0 )·sin δ

and the stress can be written as:

σ = ε0·(E1·sin ωt + E2 ·cos ωt)

tan δ = E2 /E1

Since E 1 is the elastic part of the modulus, it is called the “storage modulus”. It is,
namely, a measure for the energy stored in a reversible way. E 2 denotes the viscous
part, and is thus a measure for the energy dissipated into heat per period of vibration;
Visco-elasticity 111

it is called “loss modulus”.

Periodic vibrations, mechanical as well as electrical, are often represented as a

uniform motion along a circle with angular velocity ω . The projection on a
horizontal axis carries out a harmonic vibration r·cos ω t (Figure 6.10). This can,
mathematically, be expressed in a very elegant way by considering the rotating point
as a complex number:

r* = x + iy (i = 
x is then the projection on the horizontal axis, varying according to x = r·cos ωt, and
is the real part of the complex number r*; iy is the imaginary part. This way of
representation is based on the well-known formula:

eix = cos x + i·sin x


r r cos ωt

Figure 6.10. Complex representation of a harmonic vibration.

A periodic deformation with amplitude ε0 is then given by:

ε = ε0·exp(iωt)

while we tacitly assume that we consider the real part, ε = ε0·cos ω t, only. The
expression for the stress then becomes:

σ = σ0·exp i(ωt + δ)

(by which we mean its real part:

σ = σ0·cos(ωt + δ) )

The phase difference δ is, in the complex plane, the angle between the radius vectors
of σ and ε . The E modulus can now easily be written as a complex quantity, E*
(Figure 6.11):
112 From polymers to plastics

E2 E*

Figure 6.11. Complex modulus of elasticity.

σ σ0·exp i(ω t + δ) σ0 σ0
E* = = = ·exp(iδ) = ·(cos δ + i·sin δ) =
ε ε0·exp iω t ε0 ε0

σ0 σ0
= cos δ + i sin δ = E1 + i·E2
ε0 ε0

In contrast to σ and ε , E* retains its nature as a complex quantity, composed of a

storage modulus (the real part) and a loss modulus (the imaginary part).

The use of complex numbers appears to offer considerable advantages in simplifying

calculations on vibrating systems. A damped vibration, for instance, is represented

ε = ε0· exp(iω t) ·exp(–αt) = ε0· exp(iω – α)t

This expression is much easier to handle in differentiation etc., since the periodic
part, exp(i ω t) (which means: cos ω t) and the damping part exp(–α t) are combined
into a single exponential function.

Various types of vibration experiments can be carried out to measure E 1 and E 2 at a

certain frequency. An example is the torsion pendulum, in which the sample,
connected to an auxiliary mass, is brought into a free torsional oscillation. From the
frequency of the pendulum (around 1 to 10 sec) E 1 is calculated, from the rate of
damping tan δ and E 2 . Other types of dynamic mechanical measurements can be
carried out at higher frequencies, such as bending vibrations with or without extra
mass, wave propagation, etc. By combining a number of these different techniques, a
time scale ranging from 10 to 10-8 sec can be covered.

6.3. Integration
In combination with creep and relaxation measurements a total time span of ≈ 10-8
sec to 108 sec (3 years) can be covered.The usual techniques to measure the response
of a polymer over this time span are schematically represented in Figure 6.12.
Visco-elasticity 113

frequencies forced vibrations
sound free creep and relaxation
waves vibrations

–8 –6 –4 –2 2 4 6
10 10 10 10 1 10 10 10 sec
1/ω , t

Figure 6.12. Time and frequency span for various methods.

When we try to combine the results of these various methods into a single integrated
picture, we have to take into account three points:

– With some experiments a stress is imposed, such as with creep tests and with
vibrations in which the sample is subjected to a periodic stress. With other
experiments the deformation is given and the resulting stress is being measured,
such as with stress relaxation and with vibration tests with an imposed periodic
strain. The results of these two different types of methods cannot directly be
translated into each other but only by complex transformations, in other words:
D(t) ≠ E(t)

– In various experiments different elastic constants are being determined; with a

torsion pendulum, for instance, the shear modulus, G, is measured, with creep or
vibrations in elongation or in bending the Young’s modulus, (tensile modulus), E.
For an isotropic material the relation between E and G is as follows:

E = 2·G·(1 + ν)

in which ν is the Poisson ratio, or the ratio between transverse contraction and
elongation in uniaxial strain. For ideal rubbers (no volume change upon
elongation), ν = 0.5; for glassy polymers and polymer crystals ν ≈ 0.33, so that
E/G is in between 3 and 2.7. Complications arise when the transverse contraction
in one direction is hindered, for instance with forced vibrations of a sheet; in such
a case not E is measured but E/(1 – ν2). Moreover, the Poisson ratio also depends
on time or frequency.

– In the third place the question arises how the time and frequency scales have to
be joined together to connect results of creep or relaxation to those of vibration
experiments. It will be clear that here the superposition principle can be applied,
since with vibrations the imposed stress (or the strain) varies continuously. Using
this principle and using the complex E modulus (or the complex compliance D), it
can be shown that, with a few simplifications, t and 1/ω can be considered as the
same parameter, while ω = 2πν (ν is the frequency in cycles/sec). It has been
114 From polymers to plastics

shown that in this way static and dynamic experimental results can be joined
together satisfactorily.

Many measurements have been carried out with a large number of, sometimes very
complicated, techniques. A shorter route towards a global knowledge over such a
broad range of time and frequency, is supplied by the overall equivalence of time and
temperature. In its simplest form, this equivalence can be expressed via the
assumption that the effect of temperature on a molecular process often follows
Arrhenius’ law:

τ = c·exp(A/RT)

in which τ is the relaxation time of such a process. Also the frequently used model
representations with springs and dashpots lead to this equation: If we suppose the
spring constants, E, to be independent of temperature, and the viscosities η as
η = c′ exp(U/kT) (which is the usual behaviour of a viscosity), then the relaxation
time is

η c′
τ = E = E exp(U/kt) (= c·exp(A/RT))

With an increase in temperature from T0 to T the modulus curve will then be shifted
along the log t axis over a distance:
A 1 1
R ·( T – T 0 )
ln aT =

in which the quantity aT is defined as:

aT = t (T ) (Figure 6.13).
rel 0

log E

T0 T < T0
log aT

log t

Figure 6.13. Time-temperature shift.

This offers the possibility to carry out measurements in a limited interval of log t (of
log ω ) at several levels of temperature, to derive the shift factor a T (T) from the
results, and to construct a master curve over a broad range of log ω . This is
Visco-elasticity 115

illustrated in Figure 6.14.

log E


T5 T6



1 sec
–8 8
10 10

Figure 6.14. ‘Mastercurve’.

Along the routes described thus far, it is possible to characterize E and its both
components E1 and E 2 , either in the complete, elaborate way or via the short-cut
using the t-T relation, over a broad region of log t (log ω ). A general schematic
picture of E 1 and E 2 as functions of frequency or time is given in Figure 6.15.
Generally speaking, two maxima in E2 occur, corresponding to two transitions in E1 .
Next to these main transitions, a number of secondary transitions may be present.
These are, however, not adequately represented in the resulting curve from a log t –
T shift, which provides, therefore, despite of its simplicity, a very limited overall
view only.

log E



log t log ω

Figure 6.15. Two relaxation mechanisms. Figure 6.16. Two (broad) Maxwell

The pattern in Figure 6.15, read from the right to the left on the log t scale, resembles
the relaxation behaviour, the decrease of E with increasing time under a constant
strain, for two Maxwell elements in parallel (Figure 6.16), though over a much
broader interval of log t. It could, therefore, be considered as the behaviour of two
broad clusters of Maxwell elements. The log E –log ω curve thus indicates the
existence of two broad relaxation mechanisms, each with a spectrum of relaxation
116 From polymers to plastics

times. This general picture is shown by all polymers, except by permanent networks,
where the left-hand (low ω, long t) mechanism is not present.

After what has been said about the T-t equivalence, it is not surprising that the time
dependency of E1 resembles the T-dependency, which we have considered in detail
before. In E(log t) we see, indeed, the same phases and transitions as in E(T) (Figure
6.17). It should be remarked that this time-temperature equivalence only holds for
amorphous polymers or for the amorphous part in semi-crystalline polymers.

log E1 log E
tan δ tan δ



log ω log t Tg T

Figure 6.17. Comparison of time- and temperature scale.

In practice, E(T) can be determined in a much easier way than E(t); it can be done
with a single device, e.g. a torsional pendulum, or its modern equivalent, DMTA, at
an about constant frequency. The glass-rubber transition covers some tens of degrees
on the T-scale; on the time scale it extends over many decades.

Both E2 and tan δ show maxima in the regions where E1 strongly changes. This can
be understood as follows: At very high frequencies, thus in the glassy region, the
chain motions are so small that they hardly give rise to energy losses (in spring-
dashpot models: the dashpot hardly moves and the whole deformation is taken up by
the spring). At very low frequencies, e.g. in the rubbery region, losses also hardly
occur; though the viscosity plays a role, the rates of deformation are so small that the
energy dissipation is low. In between, in the transition region, these extreme
situations do not exist, and the damping, the energy dissipated per cycle, shows a

Mechanical properties

7.1. Stress-strain diagram

A simple representation of a number of mechanical properties is given by the stress-
strain diagram. Such a diagram can be produced in a tensile test, in which the force
is recorded as a function of the elongation, which increases at a constant rate. The
force is divided by the initial cross-sectional area, and the strain by the initial length,
so that a stress-strain curve is obtained, which is independent of the geometry of the
sample. Besides in elongation, the sample can also be deformed in bending, in
compression or in torsion; this leads to different diagrams which are, as a matter of
fact, related to each other.

A schematic example of a stress-strain diagram is given in Figure 7.1. From this

curve the following properties can be read off: The modulus of elasticity is the slope
of the first, approximately straight part of the curve. Here the relation E = σ / ε is
valid, or, with an initially curved line, E0 = (d σ/dε) for ε = 0. To define the modulus
at a higher value of strain, ε1, the secant modulus E s = σ1/ε 1 can be taken, or the
tangent modulus Et = (dσ/dε) for ε = ε1 (see Qu. 7.1).


σy Etan

ε1 εy εb

Figure 7.1. Stress-strain diagram.

Often the material shows a yield point at a certain stress, the yield stress σy; the
stress increases only slowly from this point or may even decrease. In fact the
material already fails at this yield stress; actual fracture occurs at a somewhat higher
stress σb and a considerably higher strain εb. The total area under the curve is the
work required per unit volume to break the material, and is a measure for its
118 From polymers to plastics


We can rank the polymers according to the levels of the various properties: stiff or
soft, tough or brittle, strong or weak plastics (see Figure 7.2).

The forces are expressed in newton (N), the stresses and moduli of elasticity in N/m2
= Pa (pascal), also in MPa = 106 N/m2 = N/mm2 or in GPa = 109 Pa.
σ σ brittle (e.g. PMMA) σ strong (e.g. PVC)
stiff (e.g. POM)

tough (e.g. PC) weak (e.g. ABS)

soft (e.g. LDPE)

ε ε ε

Figure 7.2 Various types of stress-strain diagrams.

7.2. Stiffness and creep

7.2.1. Modulus of elasticity
In general the stiffness of plastics is lower than that of other materials. Figure 7.3
gives (left) a schematical survey of the moduli of a number of materials. As a matter
of fact, rubbers are the lowest. The thermoplastics show a broad spectrum of moduli,
ranging from the softer types as PE, PP, polyamides, fluor polymers and cellulose
plastics, up to a bigger group of harder ones in the region of 2 to 4 GPa. Thermosets
are slightly higher in stiffness.

The gap between rubbers and the softer thermoplastics is being bridged more and
more by the thermoplastic elastomers, TPE’s (not indicated in Figure 7.3, but later.
in Figure 7.27); the stiffness of these block copolymers can be adjusted by the choice
of block type and block length ratio.

Plastics compare more favourably with other materials if the stiffness per unit mass
is considered, or the modulus of elasticity divided by the specific mass. This is
demonstrated by the right hand side of Figure 7.3; as a result of the lower density of
polymers the values are closer to each other; an advanced composite now amply
exceeds steel!

In comparing the E-values of the various polymers, the best starting point is formed
by amorphous polymers in the glassy state; for this group E is mostly about 3 to 3.2
GPa (e.g. PVC, PS, PMMA), except when, below room temperature, a secondary
transition occurs, such as with PC E ≈ 2.1 GPa).
Mechanical properties 119

Higher values can be reached for semi-crystalline polymers below Tg ; the crystalline
phase is stiffer than the glassy amorphous phase (e.g. PEEK, E ≈ 4 GPa). Semi-
crystalline polymers above Tg have, however, a much lower E-value, such as PE
(0.15 to 1.4 GPa) and PP (≈ 1.3 GPa); E is, in these cases, strongly dependent on the
degree of crystallinity and on the distance to Tg. Sometimes a low modulus is also
found for semi-crystalline polymers below Tg , due to the effect of one or more
secondary transitions; a strong example is PTFE (E ≈ 0.6 GPa!).

soft thermoplastics
hard thermoplastics
textile fibres
thermoplasts with glass fibres
thermosets with fillers
thermosets with glass reinforcements
thermosets with C-fibres
1 10 10 2 10 3 10 4 10 5 10 6 1 10 10 2 10 3 10 4 10 5 10 6
E (MPa) (
E/d MPa / kg 3 )

Figure 7.3. Absolute and reduced moduli of elasticity.

Higher values for E also occur with strongly cross-linked systems (thermosets); an
example is an unfilled epoxy resin with an E- modulus higher than 4 GPa.

A special category is formed by oriented polymers, which have considerably higher

stiffnesses. The most obvious example is the textile fibre; the orientation, frozen-in
in a crystalline structure, raises E by a factor of 3 to 5. Extremely high orientations,
as met in liquid-crystal polymers (LCP’s) result in even higher E-values, namely 60
to 120 GPa!

A higher stiffness is also obtained by the incorporation of hard particles or short

fibres. Particles are responsible for a 2 to 2.5 fold increase in E, while with short
glass fibres a factor of 3 to 5 can be obtained. Long fibres, forming a continuous
reinforcing phase, produce a much stronger effect; here the fibres practically carry
the whole stress, while the matrix polymer has hardly any influence on the stiffness.

Finally, a few factors should be mentioned which do not influence the modulus. The
120 From polymers to plastics

first is the chain length (molar mass) and its distribution. E is independent of M,
unless, with semi-crystalline polymers, longer chains are responsible for a somewhat
lower degree of crystallinity (e.g. with ultra-high molecular PE); in such a case E
slightly decreases with increasing M.

E is also independent of chain stiffness and chain interactions; these factors play a
role in the height of the glass-rubber transition temperature and the melting point. A
stiffer chain, therefore, does not result in a stiffer polymer; except, sometimes, in an
indirect way, namely when stiff chains enable the formation of high orientation, such
as in liquid-crystalline polymers (see § 4.6).

7.2.2. Creep
We should realize that the values of the moduli of elasticity, as discussed so far, are
only applicable to short-term loading situations. The creep, already mentioned
several times before, renders these values unsuitable to characterize the behaviour of
a polymer under stress over a longer time. In § 4.5 we already met the example of
two polymers, where POM, at a certain stress, initially deforms less than PC, but,
later on, its deformation exceeds that of PC.

The usual way in which the deformation changes with time, has been dealt with in §
6.1. The best representation appeared to be a Maxwell element with a Kelvin-Voigt
element in series; the deformation is then composed of three components: an
immediate elastic strain, which recovers spontaneously after removal of the load, a
delayed elastic strain which gradually recovers, and a permanent strain. Moreover,
we noticed that a single retardation time (a single Kelvin-Voigt element) is not
sufficient: we need to introduce a spectrum:

D(t) = ∫ L(τ)·[1 – exp(–t/τ)] dτ

Creep curves can also be represented by empirical equations, such as

D(t) = a + b·t1/3 (Andrade)

D(t) = D0·exp(t/t0)m (Kohlrausch, 1866, glass)

The second equation appears to be applicable to a number of glassy polymers, and

also to other materials; the exponent m is always about 3 , so that creep can be
described by two parameters, D 0 and t0 , while the immediate elastic deformation is
also taken into account (D0). As a matter of fact, D0 and t0 are temperature
dependent. When the experimentally found creep curves are shifted along the
horizontal axis and (slightly) along the vertical axis, they can be made to coincide
Mechanical properties 121

into a mastercurve, which signifies the validity of a time-temperature superposition

principle (see also Qu. 7.10).

An example of such a shift is given in Figure 7.4 as creep curves of PVC, measured
up to 1000 sec at temperatures ranging from 20° to 70 °C (Struik). The curves can
easily be shifted into a mastercurve, shown in the figure at 20 °C. It is tempting to
use this mastercurve for predicting the long-term creep; in § 7.2.4 we shall, however,
see that this is not allowed!

Figure 7.4 Creep of PVC as a function of temperature (both figures taken from L.C.E.Struik,
thesis, Delft 1977).

The analogy in the creep behaviour of various (glassy) polymers and other
substances is illustrated in Figure 7.5; on each material measurements have been
carried out over a broad range of temperatures, and all results coincide, after shifting,
into a master curve with the equation:

D(t) = D0·exp(t/t0)1/3

For practical use, this representation with two parameters is, however, not sufficient,
due to other complications in creep behaviour, namely:

– non-linearity
– physical ageing,

In the following paragraphs these complications will be discussed.

122 From polymers to plastics

Figure 7.5. Creep mastercurve for a number of materials.

7.2.3. Non-linearity
The reduced creep curves in Figure 7.6 (ε/σ or D(t)) should coincide if Boltzmann’s
superposition principle would hold; for each level of stress, D(t) should be the same.
However, creep behaviour is non-linear; linearity only occurs at very low values of
σ and ε, as a limiting case. Therefore, D(t) is D(t,σ).
10 –10 σ, MPa
10 m2/N 20
8 10
ε/σ PVC 3
1, 0,1

10 10 3 10 5 10 7
t (sec)

Figure 7.6. Non-linearity of creep for PVC.

Various models have been proposed to account for the non-linear creep of polymers,
such as the Eyring model; this model “explains” some aspects, but fails with other
ones, and will not be discussed here.
Mechanical properties 123

For practical applications empirically determined creep data are being used, such as
D(t) or, more often, E(t) curves at various levels of stress and temperature. The most
often used way of representing creep data is, however, the bundle of creep
isochrones, derived from actual creep curves by intersecting them with lines of
constant (log) time (see Figure 7.7). These σ -ε -curves should be carefully
distinguished from the stress-strain diagram discussed before, as generated in a
simple tensile test!

log time

σ1 σ2 σ3 σ4 σ5


t1 t2
σ5 t3
σ1 isochrones

Figure 7.7. Construction of creep isochrones.

From a bundle of creep isochrones the tendency of a material to creep, can be read-
off in a glance, namely from their mutual distance. Along horizontal lines of
constant stress, the increase of deformation by creep can be detected. When the
isochrones are straight lines, the superposition principle holds. Compliances at
certain combinations of σ and t follow from the (reciprocal) slopes of connection
lines to the origin.

Creep isochrones are, sometimes, used to obtain information on stress relaxation;

the stresses are then read-off on a vertical line (constant strain). In general this is,
however, not allowable, since E(t) in relaxation is not equal to 1/D(t) in creep. In a
linear region this objection is not too stringent; for want of something better, the
procedure can be used as a first approximation (data on stress relaxation are very
124 From polymers to plastics

1 10 102 103 104 1 10 102
20 hours 20
σ σ 103
15 15
10 PPO/PS 10
(Noryl) hours
5 23 °C 5 PBTP
23 °C
ε ε
0 % 1.5 0 % 1.5
0.5 1.0 0.5 1.0

Figure 7.8. Examples of creep isochrones.

Figure 7.8 gives some bundles of isochrones as supplied by a polymer manufacturer

(GE). As we saw before, when comparing POM with PC, also here the higher rate of
creep of a semi-crystalline polymer (PBTP), compared with the amorphous blend of
PPE with PS, is obvious.
σ100 σ80 σ60 σ40 σ20 –2
10 1
10 2

10 4


100° 80° 60° 40° 20° C ε

Figure 7.9. Creep isochrones at different temperatures.

Creep at different temperatures can be represented by separate bundles of

isochrones. A simple time-temperature shift could mean that for higher temperatures
shorter time values could be written at each curve. It is, however, easier to transform
the stress scale. Sometimes this is possible with sufficient accuracy; in those cases
only one bundle of isochrones is given with different stress scales for a number of
temperature levels (Figure 7.9) (see also Qu. 7.13).

7.2.4. Physical ageing

In § 3.3 the effect of cooling rate on the free volume in the glassy state has been
discussed. Rapidly cooled glassy polymers have a greater free volume, but they
show volume retardation. This volume change, though very small, has a
considerable effect on the creep behaviour: all relaxation times for creep are shifted
towards higher values. This phenomenon has been studied extensively by Struik
(thesis Delft 1977).
Mechanical properties 125

An important consequence of physical ageing is that, during a creep experiment or

during creep in practice, the rate of creep is continuously reduced. Quantitative
theories, supported by experimental results, give the picture as presented in Figure
7.10: the dotted lines indicate the creep as extrapolated from observations at higher
temperatures; the drawn lines are experimentally determined. The differences, in a
favourable sense, are enormous!

Figure 7.10. Effect of physical ageing on creep.

These large differences clearly demonstrate the impossibility to apply time-

temperature extrapolations in an ageing glassy polymer. Ageing proceeds, at a
temperature below Tg – 25°C, over thousands of years, and is, therefore, an essential
factor when a polymer is subjected to long-term loading. If the physical ageing
would not cause a continuing increase in the resistance to creep, plastics would not
be able to withstand mechanical loading over longer times: the creep would go on
and on, since its upper limit is the rubber compliance!

Creep data should, therefore, be considered with care. In most cases the strongest
ageing effects are already taken into account more or less automatically.

Summarizing: The basic idea, mentioned in chapter 6, that creep of solid polymers
could be represented by a simple four-parameter model (the Burgers model),
composed of a Maxwell and a Kelvin-Voigt model in series, appears to be
inadequate for three reasons:

First, already mentioned in Chapter 6, a single retardation time is not sufficient to

account for the real creep behaviour, which extends over many decades on the time
scale; we need a spectrum of retardation times.

Secondly, it appears that a spectrum is also insufficient, since this is based on linear
126 From polymers to plastics

viscoelastic behaviour; in reality creep is practically always non-linear. The same

objection holds for other, empirical, equations, such as the one of Kohlrausch.

In the third place, it should be taken into account that the material, before and during
creeping, is not constant: physical ageing occurs, which exerts an important
influence on the creep.
σ σ

ε ε

a. single element b. spectrum (linear)

σ σ

ε ε

c. non-linear d. with physical ageing

Figure 7.11. Three complications with the representation of creep by a single linear

These three complications are schematically shown in Figure 7.11, using creep
isochrones; as a reference (a) a Kelvin-Voigt element has been chosen with a spring
in series (a Burgers model without irreversible flow).

7.3. Damping
In Chapter 6 we have seen that in a strained polymer a part of the stress is used to
store energy (the elastic part), while the remaining part is dissipated into heat (the
viscous effect). This can, in a very simple way, be demonstrated by a stress-strain
diagram with increasing and decreasing strain (Figure 7.12). The work required to
strain the polymer from ε = 0 to ε = ε1 is the area below the upper curve, while the
recovered energy when ε decreases to zero is given by the area below the lower
curve; these energies can, respectively, be expressed as:
Mechanical properties 127

ε1 0

∫ σ·dε and ∫ σ·dε

0 ε1

The difference between these quantities, the indicated area, is the dissipated energy,
also called the hysteresis work.


Figure 7.12. Hysteresis.

The two contributions to the deformation energy can be expressed by the quantities:

E* = E1 + i·E2 the complex modulus of elasticity

E1 the storage modulus
E2 the loss modulus
tan δ = E2 /E1 (see § 6.2)

In Chapter 6 we have seen that tan δ, a measure of the relative energy dissipation,
depends on temperature and frequency, and that it shows maxima at transitions. A
strong maximum occurs at the glass - rubber transition, weaker maxima at secondary
transitions. In general: tan δ is higher when the E(T) curve is steeper.

The loss factor, tan δ , can be measured with the aid of dynamic-mechanical
experiments (such as the torsion pendulum). The deformation in such a test varies as
indicated in Figure 7.13; the damping follows from the “logarithmic decrement”, Λ;
it can be easily shown that

Λ εn
tan δ =
π with Λ = ln εn+1

An attractive simple experiment to demonstrate the damping is rotating bending:

take a thin round rod, bend it, and rotate it in the bent condition without applying a
torsion; then a clearly detectable resistance is met. The torque required for rotation
is, according to a simple calculation, proportional to E2 .

In several respects damping is important for practical properties:

128 From polymers to plastics

εn εn + 1


Figure 7.13. Damping factor in a free vibration.

- If it occurs at certain temperatures and frequencies it may contribute to a better

impact strength.
- It is, of course, a predominant property in vibration damping devices.
- Damping is responsible for heat dissipation under fatigue conditions: the
temperature increase may cause drastic changes in the material properties.
- The energy consumption during repeated deformation, e.g. in tyres, is governed
by the damping.
- The friction of a tyre on the road is also largely dependent on the damping
properties of the rubber.
Since the latter two examples lead to contradicting requirements, we shall consider
the damping behaviour of a automotive tyre in more detail. For that purpose it is
useful to look first at two extreme cases: butyl rubber and butadiene rubber. The
former has an extremely high damping; the rebound of a falling ball is not more than
10 or 20 %, whereas with the latter 80 or 90 % is reached. A car tyre with a tread of
butyl rubber would have an exceptionally good grip on the road, but it attains, in a
very short time, an intolerably high temperature. A BR tread, on the contrary, is
hardly heated up, but renders the car practically unmanageable. By blending with
other rubbers (NR or SBR) a compromise can be reached.

Nearly always SBR is used in tyre treads for passenger cars; this rubber shows a
favourable balance between heat-build-up (h.b.u.) and grip. Continuing efforts are
being spent to obtain further improvements in both directions, which, on first sight,
seems to imply a contradiction. The key to reach an improvement is the fact that the
two damping mechanisms take place at strongly different frequencies; the heat
generation is governed by the damping at the rotation frequency of the wheel, but for
the friction much higher frequencies are relevant, namely those at which minute
volume elements vibrate when in touch with the road surface.

The problem is, therefore, to influence the damping in these two frequency ranges
independently of each other. Because time and temperature have analogous effects,
we consider the diagram of tan δ as a function of T for a vulcanized rubber (Figure
7.14). The frequencies for friction are so high that we already approach the peak at
Mechanical properties 129

the glass-rubber transition, Tg . Tg can be chosen at will; for a normal SBR it is about
–65 °C; an SBR with a higher styrene content shows a higher Tg and thus a higher
damping and a better friction. However, the shift of the peak also results in a higher
h.b.u.. The magnitude of tan δ at temperatures considerably higher than Tg , is hardly
dependent on T, but is mainly governed by the perfection of the network, in
particular by the number of loose chain ends. These do not contribute to the
coherence of the network, but their free mobility gives rise to damping (see also
§ 2.7).

tan δ


log f Tg T

Figure 7.14. Damping curve for a rubber vulcanisate.

Both targets could thus be reached by a slight increase of Tg (by a higher styrene
fraction), as well as a reduction of the number of chain ends. The latter can, in
principle, be accomplished by:

– higher molar mass, if allowable for the processability;

– narrower molar mass distribution (higher Mn );
– chains without any branching;
– end groups with some interaction with each other;
– end groups which specifically react with sulphur.

The damping curve B in Figure 7.15 then shows both improvements in comparison
with A.

Recent developments are in another direction, namely in the creation of different

glass-rubber transitions by the formation of blocks with different styrene content and
different sterical structure of the butadiene chain parts.
130 From polymers to plastics

Tg v
tan δ

less loose
A B ends

lower heat- B
higher build-up
log f T

Figure 7.15. Improvement in grip and in heat-build-up.

7.4. Strength
7.4.1. Tensile strength and ultimate strain
An important mechanical property is the strength, which is the stress at which the
material breaks. Besides, the strain at break is of relevance, since this indicates
whether the material is brittle or tough. Figure 7.16 gives a survey of both properties
for a number of polymeric materials in comparison with other ones.

It appears that the levels of tensile strength of polymers are less widely different than
the moduli of elasticity; the softer plastics are, together with the rubbers, all within
the range of about 15 to 30 MPa; the harder ones are between 30 and 80 MPa. For
reinforced resins the strength values are of the order of magnitude as those for fibres
and metals. The diagram further shows that thermosets, PS, PVC and PMMA belong
to the brittle polymers, while other ones, such as PA, PP,PE and PTFE are very
ductile. It should, however, be remarked,that a shift in temperature or on the time
scale may change the picture considerably; PP, for example, behaves in an impact
test at 0 °C as a brittle material, despite its high strain at break under standard

Just as the stiffness, also the strength of polymers is strongly dependent on the time
scale of loading. We distinguish two cases, namely long-term loading and impact
loading, which we shall discuss in the following sections.
Mechanical properties 131

tensile strength
500 2

200 5

100 7 8 16
6 11
9 12 17
50 10 13 18 19
14 20 23
15 21 24
20 22

0.2 0.5 1 2 5 10 20 50 100 200 500 1000 %
strain at break

1. reinforced thermosets 9. PMMA 17. PPO, PC

2. cotton 10. PS 18. PETP
3. construction metals 11. PVC 19. PA
4. synthetic fibres 12. POM 20. PP
5. wool 13. CA 21. HDPE
6. thermosets 14. TPS 22. LDPE
7. PPS 15. ABS 23. PTFE
8. SAN 16. PSU 24. rubber

Figure 7.16. Tensile strength and strain at break of various materials.

7.4.2. Long term strength

The effect of the time of loading on the breaking stress can be visualized in a simple
way. If we carry out creep tests on a number of samples at different levels of stress,
then some of these tests, at the higher stresses, will end with rupture after a minute or
after a year, as indicated in Figure 7.17. From the stress levels and the times to
failure, Figure 7.18 can now be constructed, in which also the times to attain a
certain strain have been sketched (isometric creep curves). We are most interested in
the breaking curve, which gives the relation between stress applied and and the time
to failure, or the dependence of tensile strength on the time of loading, for ductile
failure. For practical use, this curve could be extrapolated along the log t-axis to e.g.
50 years, in order to estimate a “safe” stress level for use.

There is, however, a complication: besides ductile fracture also brittle failure may
occur, which is not brought about by deformation and flow of the material, but by
the initiation and propagation of brittle cracks. Since both processes also need time
to develop, the time to brittle failure can also be represented by a curve, similar to
132 From polymers to plastics

the one in Figure 7.18. Because the breaking mechanism is now totally different
from the one in ductile failure, the position and the shape of this curve are governed
by different factors; it could have the shape of the line b in Figure 7.19, which
intersects the curve a for ductile failure at two points.

σ4 rupture

log t

Figure 7.17. Creep to failure at various stress levels.

log σ


log t

Figure 7.18. Time to failure or to a certain strain.

When a plastics material shows this picture, we can distinguish between three
regions of the stress σ : In the region with the highest stress levels, A, the applied
stress will result in a brittle fracture after a very short time. At lower stresses (B), the
time to ductile failure is shorter than the time to brittle failure, so that the material
breaks in a ductile manner. At even lower stresses (region C), again brittle failure
occurs, but now after very long times of loading.

This situation indeed occurs frequently, for instance with PVC pipes, where high
internal pressures give rise to explosive brittle fractures, while with lower pressures,
after a shorter or a longer time, balloon-like fractures occur. At even lower pressures
brittle hair cracks are formed, also with PE, causing the pipes to leak. This happens
after very long times of use (e.g. some years), but considerably earlier than predicted
from extrapolation of the curve for ductile failure.
Mechanical properties 133

log σ A


log t

Figure 7.19. Two types of failure.

With such a complex failure behaviour, it is difficult to make a reliable estimation of

the allowable stress, i.e. the stress at which, in the required duration of use (mostly
50 years for pipes for transportation of water), no failure will occur. Evidently,
extrapolation of failure tests, even if they extend over a year, is insufficient. The
transition from ductile to brittle crack failure, and the accompanying change in slope
of the curve, may take place after several years! One, therefore, resorts to tests at
elevated temperatures, on the basis of the idea that a temperature increase will
accelerate both failure mechanisms, and also the transition from ductile to brittle

Figure 7.20 gives an example of this procedure. From measurements of the time to
failure at various levels of stress and of temperature, two-dimensional extrapolation
allows an estimation of the position of the brittle failure curve at the temperature of
use. Needless to say that this a very time-consuming procedure; especially when we
realize that the scatter in measured times to failure (up to a factor of ten between the
extremes) necessitates five to ten tests to be carried out at each condition (see also
Qu. 7.14).

This procedure is mainly applied to plastics pipes. The reason is, that these form one
of the rare (but large-scale!) examples that during a long time of loading no fracture
is allowed to occur. With other applications of plastics in load-bearing constructions
the ultimate strength plays hardly any role, since in most cases the strain, and thus
the creep, is the limiting factor.

A further question is, how to act when the time to crack failure is too short to allow
for a certain long-term application? Obviously, a grade with a higher molar mass
should then be chosen, since the position of the curve for brittle failure is strongly
dependent on the chain length (contrary to the one for ductile failure, which is
largely independent of M) (see Figure 7.21). Apparently the number average molar

mass, Mn , is of importance for brittle failure, since the number of chain ends is the
governing factor.
134 From polymers to plastics

σ 20 °C
measured extrapolated
20 ductile
35° HDPE
80° expected stress
to rupture
95° (20 °C; 50 years)

safe working
5 stress

time to failure
–2 2 4 6
10 1 10 10 10 hours

1 10 100 years

Figure 7.20. Extrapolation procedure for HDPE pipes.

log stress
ductile brittle
higher M n

surface active
log time to failure

Figure 7.21. Effect of molar mass and of environment on the time to brittle failure.

The environment also plays a role; in some environments brittle crack failure is
strongly promoted. For example, detergents such as synthetic soaps can decrease the
time to brittle failure of PE by a factor between 10 and 50 (see Figure 7.21). This
phenomenon is known as stress corrosion or environmental stress cracking (ESC)
(see further § 8.5).

Long before a brittle hair crack has come to its complete development, it is already
present as a nucleus, and the material has already undergone irreversible damage.
Microscopic cracks are often partially filled with material in the form of very thin
fibres, forming bridges between the fracture surfaces. These cracks are denoted as
Mechanical properties 135

Figure 7.22. Schematic representation of a craze.

crazes; such a craze is schematically shown in Figure 7.22.

Detection of crazes is very difficult; yet craze formation should be avoided in critical
applications (such as gas pipes). In order to achieve that, one uses the expression
“critical strain”, which is based on the observation that, as long as a certain level of
strain is not exceeded, no crazes are formed. It would be convenient if this critical
strain would be independent of the applied stress; this is, unfortunately, not the case,
as illustrated in the bundle of isochrones in Figure 7.23

no t1 craze-
damage formation



ε cr ε

Figure 7.23. Bordering curve for the formation of craze.

For practical purposes the lower limit of the critical strain can be used as a criterion.
This value appears, moreover, to be not much dependent on temperature, so that it
can be considered as a material constant. It varies from ≈ 0.3 % for PS to ≈ 2.2 % for
PP. From creep isochrones a stress level can now quite easily be detected at which,
within a given time of usage, no damage to the material is to be expected. This stress
level is, of course, much lower than the one we found from Figure 7.20.
136 From polymers to plastics

7.4.3. Impact strength

For many applications the resistance of the material to shock loading is an important
property. Here we find ourselves, contrary to the previous section, on the very short
side of the time scale. For the impact strength the short-term tensile strength as well
as the ultimate strain play a role; the impact strength is, in fact, the energy needed for
rupture at a high rate of deformation, ∫ σ·dε. It is very difficult to characterize the
impact strength accurately and uniquely.

In most cases the impact strength is determined with the aid of a standard impact
test, in which a test bar is hit by a falling hammer, after which the remaining energy
of the hammer is measured. Very often a notch has been machined in the test bar.
The stress concentration round the tip of this notch causes the fracture to be initiated
at a well-defined location and not at arbitrary surface cracks or inhomogenieties.
Moreover, the rate of deformation is, at the tip of the notch, much higher than the
average value in the test bar, so that the experiment takes place at a shorter time-
scale. On the basis of the equivalence of temperature and time, this means that the
value found belongs to a lower temperature, where the material is, in general, more
brittle; the notch more or less forces the material to fail in a brittle way, so that the
notched impact test provides a better indication of the material behaviour under
extreme conditions.

Figure 7.24 and Figure 7.25 show that the effects of the sharpness of the notch and
temperature are, roughly, equivalent; polymers for which the impact strength is
strongly dependent on temperature (PVC and PA), are also more notch-sensitive
than, e.g. ABS.

Results of notched impact tests are found in various tables of material properties; a
schematic survey of a number of plastics is presented in Figure 7.26. These values
give a first impression of the ranking of the various polymers, tested under standard
conditions. With other conditions, however, the order of sequence of the various
materials may be quite different, as already appears from Figure 7.25. Moreover,
various complications arise: Internal stresses as a result of cooling and shrinkage
have a considerable effect on the impact strength. Chain orientation renders a
material anisotropic: the impact strength in the orientation direction is increased, but
is decreased in the cross direction. Reinforcing fillers in most cases bring about a
decrease of impact strength, though in some cases, particularly with very brittle
polymers or at low temperatures, glass fibres may effect an increase (see § 9.3).
Mechanical properties 137

notched impact strength

notched impact strength
30 PA
PA PC PVC PP copol

10 10
1/4 1 4 16 64 mm 0
notch radius –80 –40 0 40 80 T °C

Figure 7.24. Impact strength as a Figure 7.25. SImpact strength as a function of

function of notch tip radius. temperature.

notched impact strength

thermosets with fillers
UP and EP/glas
1 2 5 10 20 50 100 kJ/m

Figure 7.26. Impact strength of some plastics.

Glassy polymers are, mostly, brittle (e.g. PS), unless, below room temperature, a
secondary transition is present; a strong example is PC, which shows a very high
impact strength. Why such a transition is, below the temperature of use, still active,
138 From polymers to plastics

can be explained by the time-temperature equivalence: the time-scale under impact

conditions is very short!

Semi-crystalline polymers, such as PE an PP, are tough at temperatures above Tg ,

though for PP (Tg ≈ –15 °C) the critical temperature limit is about room temperature;
here also the time-temperature equivalence plays a role. Below Tg, semi-crystalline
polymers have a low impact strength (unless secondary transitions occur).

The impact strength of brittle polymers is often improved by the incorporation of

small rubber particles, either by blending or by copolymerization (e.g. PP

Finally, it should be remarked that the impact strength is also dependent on the chain

length, in particular on the number average molar mass Mn (the number of chain
ends being the governing factor).

7.5. Surface properties

7.5.1. Hardness
Hardness is the resistance of a material against penetration of a harder body. A great
number of methods exist to characterize hardness; for plastics materials these
methods have mostly been taken over from metal testing methods and some from
rubber testing. The various methods fall apart into two, very much different,
categories, namely:

– methods at which the penetration is measured during the presence of the load on
the penetrating body,
- methods where the measurement is carried out after removal of the load.

To the first category belong the α -Rockwell and the Shore A, C, and D hardnesses.
With α -Rockwell the penetrating body is a 2 inch diameter sphere, loaded with 60 kg
force; after 15 sec loading the penetration depth h (mm) is measured, and the
hardness is defined as R α = 150 – h/0.002. Practical values for plastics range from
Rα = 20 to 150.

With the Shore methods the penetrating body is a cone; with A and C the tip of the
cone is flattened to a circle with 0.79 mm diameter; with D it is rounded-off to a
radius of curvature of 0.1 mm. The force is exerted by a spring, which for A, mainly
applied to rubbers, is considerably less stiff than for C and D. Hard thermoplastics
are mostly characterized by Shore D; the values found are between 50 and 90 units.

The methods suggest that these hardness values are related to the modulus of
elasticity of the material. Round the penetrating body a complicated stress
Mechanical properties 139

distribution is present; the state of deformation is also difficult to analyse. Yet in fact
a deformation is being measured at a given stress. In Figure 7.27, for a number of
plastics together with some thermoplastic elastomers, the E-modulus and the Shore-
D hardness, as read from material tables, are plotted against each other. The
correlation appears to be satisfactory if we take into account that the deformations in
the hardness test may, locally, be rather high, and that the time scales of deformation
are different.




40 thermoplasts
thermoplastic elastomers
50 100 200 500 1000 2000 5000 MPa

Figure 7.27. Relation between hardness and E-modulus.

With the second category of hardness tests the measurement is carried out after the
load is removed. This is the case with Rockwell R, S, V. L, M and P; under a small
prestress the position of the sphere is measured, so the permanent penetration depth,
1 1
h (mm). The sphere diameter is 4 or 2 inch, the load is 60, 100 or 150 kg force,
dependent on which of the six types of test is chosen. The hardness is defined as HR
= 130 – h/0.002. This hardness value has no relation at all to the modulus of
elasticity; the permanent deformation after recovery is being measured (such a type
of test would result in a very high value for rubbers!).

Another criterion is used with the Vickers hardness test; after penetration of a
pyramid-shaped diamond under stress, the diameter of the indent is measured after
removal of the diamond. The hardness is defined as the applied force divided by the
area of the indent. This is again a measure of the permanent deformation, or,
possibly, of the yield stress.

7.5.2. Friction
Friction is an extremely complex phenomenon. The simplest fundamental treatment
starts from the assumption that the surfaces, in touch with each other, are rough on a
140 From polymers to plastics

Figure 7.28. Contact between two surfaces.

micro-scale, so that the actual contact area is only very small (Figure 7.28).

The surfaces are pressed together with a force N, and they touch each other over an
area A. The normal stress in the contact surface is thus N/A. This value can be
considered as the yield strength of the material, σy, so σy = N/A and A = N/σy.

During a sliding movement the points of contact, where the two surfaces are welded
together, are continuously being broken, whereafter they reform; this requires a force
τb·A , in which τb is the shear strength of the material. This is the friction force, F =
τb·A = µ·N, where µ is the coefficient of friction. This leads to a very simple
expression for µ:

µ = τ b /σ y

However, there are, as always, complications: both τb and σ y are functions of time
(viscoelasticity) and temperature. In particular the temperature effect is of
importance; the continuing plastic deformation results locally in high heat
dissipation, which, due to the low heat conduction of polymers, is not transported to
the environment. An estimation of the local temperature increase can be made with:

2 q ′ ·√
σyT = T0 + ·
 √λρc
in which
q′ = heat production per unit time and area
λ = heat conduction coefficient
ρ = specific mass
c = specific heat

q′ equals the heat (or work) of friction:

W·l W µ·N
q′ =
F·t = F ·v = F ·v = µ·σN·v
(F = total surface area, v = sliding speed, σN = normal pressure), so
2 µ
∆T = ·σN·v·√

Mechanical properties 141

Substitution for certain polymers and conditions easily leads to values of several tens
of °C.

For a semi-crystalline polymer, the coefficient of friction depends on temperature as

schematically indicated in Figure 7.29:

Tm T

Figure 7.29. Coefficient of friction as a function of T.

After an initial decrease, a strong increase occurs as a result of the sticking together
of the sliding surfaces. From the melting point, µ again strongly decreases; the
surfaces are then molten.

So far we discussed friction between a pair of equal polymers. When the partners are
different, e.g. a plastic against steel, the picture changes. In this case the real contact
area will also be much smaller than the apparent one, but the steel “peaks” will, not
flattened themselves, penetrate into the polymer, and will “plough” through the
polymer when the surfaces slide over each other. It appears that, also in this case, the
coefficient of friction can be expressed as µ = τb/σ y , where τ b en σ y are the values
for the polymer. A great difference is the much higher rate of heat transportation,
which changes the whole picture. It has been observed that the friction between a
metal and a polymer, after an initially high value, rapidly drops to a much lower
level. Abraded polymer particles stick to the metal surface and then cause polymer-
polymer friction, now, however, with a good heat transport.

A global survey of friction coefficients of various polymers is given in Figure 7.30

(all moving against steel). In general, polymers exhibit less friction than metals. The
lowest values for the friction coefficient are found for HDPE, POM, and, in
particular, PTFE.

The surface structure of a plastic may exert a considerable influence on the friction;
an example is given by injection moulded articles of PP, which show, in touch with
each other, a coefficient of 0.7, while for sand-blasted surfaces a value of only 0.3 is
found. For injection-moulded nylon µ = 0.65, for machined surfaces 0.47. As a
matter of fact, lubrication has a strong influence: the value of 0.47 for nylon is
reduced to 0.19 with water lubrication and to 0.08 with oil.
142 From polymers to plastics

0 0.2 0.4 0.6 µ

Figure 7.30. Coefficient of friction of some polymers.

7.5.3. Abrasion
Abrasion is also a complex phenomenon; it is not or hardly predictable from basic
properties, and is only characterized empirically by a great variety of standard

We can distinguish between, on the one hand, damage of the surface by scratches
and dents, without removal of material, and, on the other hand, damage by loss of
material. The first case is clearly related to the second category of hardness
(permanent deformation, see § 7.5.1), while the second case depends on a
combination of friction and tear strength.

The results of the numerous types of abrasion tests cannot be related to each other;
the tests are only able to compare members of the same family of materials. The
ranking in abrasion resistance, found for a series of materials, often depends on the
method chosen. Only a few significant data will, therefore, be given here. The lowest
abrasion values are found for polyamides (nylons), polyolefins (PE and PP), and
polyesters (PETP and PBTP). In particular, a HDPE grade with a very high molar
mass (up to 5 million) should be mentioned; this polymer has an unsurpassed
resistance against abrasion. (Here again, PE, in general a simple commodity
polymer, excels, just as in its electrical properties (see § 8.2.2) and as the strongest
known fibre, see § 4.6).

The relation of abrasion to friction appears from Figure 7.31, in which the abrasion
is given as a function of temperature for some crystalline polymers. Temperature
increase causes, from a certain temperature, a drastic increase in abrasion due to
sticking together of the surfaces.
Mechanical properties 143

weight loss
PE (high mol.) PA 6.6

temperature of
sliding surface
0 50 100 150 200 °C

Figure 7.31. Abrasion as a function of T for some polymers.

The abrasion resistance increases considerably when reinforcing fillers are added to
the polymer. This is, in particular, the case with rubbers: incorporation of carbon
black (e.g. 40 weight parts per 100 rubber parts) increases the life of a tyre tread
from 5,000 to 50,000 or even 100,000 km!

Further properties

8.1. Thermal properties

The temperature region in which a polymer can be used, is limited at the low as well
as at the high side. The most serious limitation at low temperatures is, as a matter of
fact, the glass - rubber transition for rubbers; below Tg they lose their rubbery nature
and pass into the glassy phase. Besides, for thermoplastics, the principle limitation is
that they become brittle at low temperatures, and thus lose their impact strength. The
other properties are not affected, but, in most cases, (E-modulus, tensile strength),
improved upon decrease of temperature. First of all we shall consider at which
temperature cold-brittleness appears.

8.1.1. Brittleness temperature

Brittleness is found with semi-crystalline polymers below their glass-rubber
transition Tg . An example is PP, which becomes brittle at about T ≈ –10 °C. PE
retains its ductile nature down to very low temperatures. Other polymers have a Tg of
some tens of °C above room temperature, such as polyamides and thermoplastic
polyesters. Various mechanisms are responsible for a reasonable impact strength at
room temperature; for polyamides this is, for instance, the absorption of water; also
secondary transitions in the glassy region may play a role.

Amorphous, glassy polymers, used far below their T g , are cold-brittle if no other
mechanisms are active; an example is PS. If a polymer has been improved in impact
strength by the addition of a rubbery phase (high-impact PS or PVC, ABS etc.), then
the cold-brittleness temperature is related to the T g of the added rubber. If the
polymer shows a secondary transition in the glassy region (such as PC), then this
governs the brittleness temperature.

The tough-brittle transition temperature is hard to define; it is, of course, strongly

dependent on the conditions, such as the time scale of the experiment, notch effects
etc. The brittleness temperature is, in general, being determined by a series of
standard impact tests, carried out at different temperatures; when 50% of the samples
are broken in a brittle way, then the brittleness temperature has been reached.
Further properties 145

8.1.2. Softening
At elevated temperatures al polymers soften, dependent on their glass-rubber
transition points, T g , and/of their melting points, T m. These temperatures limit the
practical use of plastics. To characterize the softening behaviour, in practice various
types of standard tests are being carried out, resulting in values for the “softening
temperature”, defined in different ways. The values mostly used are: the ISO Heat
Deflection Temperature (HDT) and the Vicat Softening Temperature (VST or

The ISO-HDT is based on a bending measurement on a standard test bar, loaded with
a constant force, and gradually increased in temperature; the temperature at which a
certain bending deflection is reached, is called the HDT. All conditions are
normalized: sample dimensions, rate of heating (2 °C/min), load (two cases: ISO-A
with a maximum bending stress of 1.81 MPa, ISO-B with 0.45 MPa), limit of
bending deflection (0.32 mm). The stress, σ, is fixed, as well as the deformation, ε, at
the moment of reading-off the temperature, so also E = σ/ε. In fact, the temperature
is measured at which the modulus of elasticity, E, has dropped down to a prescribed
level. These levels of E are about 1000 and 250 MPa for ISO-A and ISO-B,

With the Vicat test a steel needle, at the end flattened to an area of 1 mm2 , is pressed
into a block of the material with a standard force (1 kgf for Vicat-A, 5 kgf for Vicat-
B). The temperature is increased at a standard rate of 50 °C / hour until the needle
has penetrated 1 mm into the sample; then the Vicat softening point has been
reached. As well as we have seen with the Shore hardness (§ 7.5.1), this process of
penetration is, in fact, also governed by the E-modulus of the material, though in a
much more complicated way. Globally, also to the Vicat test a characteristic E-
modulus can be ascribed, which is lower than with the ISO bending test, namely
about 200 MPa for Vicat-B and 40 MPa for Vicat-A.

The values of the various softening temperatures can now be related to the shape and
the position of the E(T) curves, discussed in Chapters 3 and 4. A satisfactory
comparison is hardly possible, because the time scales of the tests differ too much;
E(T) curves have, in most cases, been measured by dynamical-mechanical tests at a
time scale round one second, whereas the determination of softening temperatures
extends over several minutes. In principle, however, the picture presented in Figure
8.1 is valid (see also Qu. 8.3).

In this picture the dependence of E on T is schematically sketched for three

polymers, largely differing in softening behaviour:

1. an amorphous polymer, for which E drops quite steeply in the glass-rubber

146 From polymers to plastics

log E
ISO/A conditions

Vicat conditions
amorphous crystalline
th.plast th.plast

ISO/A scale
2 1 3
Vicat scal
1 2 temperature

Figure 8.1. Softening temperatures and E(T) curves.

transition region,
2. a semi-crystalline polymer, showing a gradual drop in E between Tg and Tm,
3. a thermoset, for which E, also above Tg , has not dropped far enough to allow the
determination of a Vicat test.

From the first two examples it appears that simple usage of an arbitrary softening
temperature, taken from a table, does not provide an answer as to the temperature
resistance of a polymer: the data for Vicat and ISO-HDT contradict each other!

Apparently, the Vicat point indicates the limit where the material, without being
subjected to an appreciable stress, loses its solid-state nature, while the ISO-HDT
gives an indication of the upper temperature which can be withstood under stress. A
practical example of the discrepancy between Vicat and ISO-HDT is given in the
table below.

PC 145 °C 135 °C
PA-6 210 °C 85 °C

As a matter of fact, an E(T) curve provides much more information than the single-
point tests on softening temperatures; however, due to the time dependency of the
mechanical properties, not even enough to provide a basis for construction purposes.
For the use of plastics under load at elevated temperatures creep data are required !

Softening temperatures increase when reinforcing fillers are present or short fibres.
Tg and Tm will, of course, not change, but the increase of stiffness causes the whole
E(T) curve to shift to a higher level, e.g. by a factor of 2 for various powders (quartz,
Further properties 147

talcum etc.), and by a factor of 3 to 5 for glass fibres. The effect of reinforcement on
ISO-HDT and on Vicat is, for semi-crystalline polymers, much greater than for
amorphous glassy ones. This appears from Figure 8.2; it is, again, simply a matter of
slope of the curve!
log E log E

amorphous semi-


amorphous: ∆· ISO = 10 to 20 ° crystalline: ∆ ISO = 65 to 170 °

Figure 8.2. Effect of glass fibres on softening temperature.

The large effect on crystalline polymers is, on the one hand, trivial and self-evident;
on the other hand it appears that reinforcing fibres bring about such an increase in the
stiffness at increasing temperatures that the temperature region of application is
substantially widened.

8.1.3. Thermal expansion

Polymers show, in general, a considerably higher thermal expansion than other
materials. Figure 8.3 gives an overall survey. Roughly speaking, a relation exists
between the coefficient of thermal expansion and the reciprocal stiffness.
× 10 –5
20 soft

8 th.plasts
3 th.plasts
and aluminum
2 th.sets brass
with fillers
1 glass

Figure 8.3. Survey of expansion coefficients.

The coefficient of expansion depends on temperature, as already discussed in

148 From polymers to plastics

Chapter 3. In the glass-rubber transition the thermal expansion coefficient shows a

discontinuous jump; with crystalline polymers the coefficient increases strongly from
a few tens of °C below the melting point (Figure 8.4)
V α=


Tg T Tg T
V α=


Tg Tm T Tg Tm T

Figure 8.4. V(T) and α(T) relations.

Fillers have a strong effect on the thermal expansion; they may reduce the
coefficient by a factor of 2 to 3.

Chain orientation effects anisotropy in the expansion; in the orientation direction it

is considerably lower than across. Moreover the tendency to (irreversible) shrinkage
in the orientation direction upon temperature increase should be taken into account.

Fully reversible negative expansion is shown by a strained ideal rubber, as a result of

the entropy-elasticity, discussed in § 5.1.

8.1.4. Thermal conduction

The coefficient of thermal conduction, λ, is defined on the basis of the formula:

Q T1 – T2
= λ·F· d
in which
Q = heat flow during time t
F = area
d = thickness of the sample,
T1 and T2 the temperatures of the opposing sides.

The dimension of λ is W/m·K (Watt /meter·Kelvin).

Further properties 149

Compared to other materials such as metals, the thermal conductivity of polymers is

100 to 1000 times smaller (see Figure 8.5). Fillers increase the conduction by a
factor of 3 to 4.

natural stone
plastic foams W/mK
–2 –1 2 3
10 10 1 10 10 10

Figure 8.5. Thermal conduction of some materials

For technology the low heat conductivity is of importance as a great hinder in

processing operations. Heating of a solid mass to reach the fluid condition takes a
long time, as well as cooling down, e.g. in the mould of an injection moulding
machine. The heating process can be speeded up in several ways, namely by
dissipation of mechanical energy, utilizing the high viscosity, by high-frequency
dielectric heating or by ultrasonic sound waves; in all these cases the heat is
generated evenly throughout the whole mass (see § 11.1.1). Cooling, however, is
always dependent on the heat conduction, which, therefore, largely influences the
cycle time in processing of thermoplastics.

On the contrary, a beneficial aspect of the low heat conductivity is the possibility to
apply plastics as thermally insulating materials, in particular as foams. The heat
conduction of a polymer foam is composed of four components:

λ = λp + λg + λr + λc

λ p = conduction of the cell walls,

λ g = conduction of the gas in the cells,
λ r = heat transfer by radiation,
λ k = heat transfer by convection.

With very low foam densities λp is negligible, and λ is mainly governed by λg and by
λ r (transfer by radiation plays an increasing role with decrease in density). λ c is,
mostly, small, but may become of importance with larger cell size; within the cells
convection may play a role, which increases the heat transport of the gas in the cells
considerably. With increasing foam density the contribution of λp increases. In total,
we find a minimum in the λ-d curve, (Figure 8.6), the level of which is largely
150 From polymers to plastics

governed by the type of gas in the cells. For polyurethane foam this minimum may
be very low due to the low λg of the blowing agent, freon, though diffusion through
the cell walls will cause the thermal insulation to decrease with increasing time.


Figure 8.6, Thermal conduction through a foam as a function of density.

8.1.5. Maximum temperature of use

The most obvious limitation in the maximum temperature of use of a polymer is its
softening temperature, as discussed in § 8.1.2. Besides, degradation may play a role,
in particular when the material is exposed to high temperatures during a long time. In
various tables of material properties maximum temperatures of use are given, often
specified into short-term and long-term exposure to elevated T. An example: PC has,
as far as softening is concerned, a maximum temperature of use (ISO-HDT) of 135
°C, but due to degradation, values are given ranging from 140 °C (for hours) down to
100 °C (for years). These values are, of course, also strongly dependent on the
stabilizers which the manufacturer has added to the polymer to make it suitable for
special high-temperature applications. Often a certain polymer is available in a range
of grades with increasing resistance against degradation at elevated temperature.

The parameter most often used to indicate the maximum temperature of use is the
“temperature-index”, introduced by the American Underwriters’ Laboratories Inc.,
and, therefore, also named the “U.L. index”. Since, for different applications,
degradation affects the performance of polymers in different ways, different values
of the UL index are attributed to one and the same polymer, for instance for electrical
insulation and for impact strength.

8.1.6. Burning behaviour

All polymers burn when exposed to a flame (they are all organic substances), but
considerable differences exist between polymers in their behaviour after removal of
the flame. Some continue burning, other ones extinguish. The degree of self-
extinguishing is expressed as the “oxygen index”, which is the oxygen concentration
in the surrounding atmosphere, at which the material just extinguishes. An oxygen
concentration below 21 % means that, in a normal atmosphere, the polymer burns
on; if the index is higher, it stops burning.
Further properties 151

Some values of the oxygen index are given in Figure 8.7. PTFE (teflon) appears to
be highly self-extinguishing; other polymers containing chlorine or fluor also clearly
belong to the self-extinguishing category.

For a number of materials the burning behaviour can be improved by the addition of
bromine- or antimony compounds, mostly, however, at the cost of some mechanical
or electrical properties.

Gradually, the insight is gaining ground that the heavy smoke formation caused by
some burning polymers, involves greater risks than than their flammability itself, the
latter being not much different from large-scale applied materials such as wood,
textiles and paper. More than ever, test methods and criteria for safe use are being
developed to recognize and prevent the risks of smoke development, as regards
visibility as well as toxicity.

cotton, papier
paraffin PTFE
0 20 40 60 80 100

Figure 8.7. Oxygen index.

8.2. Electrical properties

8.2.1. Electric resistance
Plastics are excellent electric insulation materials: in their specific resistance they
exceed glass and ceramics (see Figure 8.8).

It should be remarked, however, that the specific resistance of polymers cannot

uniquely be defined; the electric resistance strongly depends on the electric tension
applied (Ohm’s law is not obeyed), and also on the time of application of the tension.
Moreover, the resistance depends on temperature; it decreases drastically upon
152 From polymers to plastics

temperature increase (by a factor of 10 or more per 50 °C). Quite often the humidity
of the surrounding atmosphere plays a role, especially if fillers are present or with
contaminated surfaces. In these cases the resistance measured is controlled by
conduction along the surface rather than through the material.
20 Ωm
10 16
PVC, PMMA, PC, EP paraffin
10 porcelain
thermosets glass
10 8 with fillers

10 4 polymers
with carbon
1 black

10 –4 graphite
10 –8 copper

Figure 8.8. Specific resistance of some materials.

As a matter of fact, the electric resistance is of major importance for electrotechnical

applications. But it also plays a role in the frequently met phenomenon of
electrostatic charging. With all kinds of materials, charges are transferred when two
bodies are brought into contact or slide against each other. A low electrical resistance
causes immediate discharge, but with good insulators the charge remains present
after breaking the contact. The high electric tensions which go accompanied with
these charges, give rise to dust attraction, sticking together of films, or even the
formation of sparks.

A rough estimation of the amount of charging can be made by considering a block of

material, with thickness d and area A, as a resistance R and a capacitor C in parallel
(Figure 8.9).

An initially present electric voltage V0 will relax with time according to:

V = V0·exp(–t/RC)
R = ρ ·(d/A) (ρ = specific resistance)
C = ε·(A/4πd)

(ε is the dielectric constant, the product of ε0, the dielectric constant in vacuum, and
Further properties 153

d ρ C R

Figure 8.9. Model for electrostatic charging.

εr, the relative dielectric constant).

The relaxation time, τ = RC, is then:

τ = ερ/4π = ε0·εr·ρ /4π

After converting electrostatic units to SI units. and assuming that εr = 2 to 4, this

leads to
τ ≈ 3·10–11 ·ρ .

If we require, for instance, that most of the transferred charge should disappear
within 1 msec (τ = 10–3 sec) , than ρ should be smaller than 3·107 Ωm. This seems a
realistic value to prevent the formation of sparks. A much less stringent requirement
is the absence of dust attraction; for that purpose τ should be lower than 100 or 1,000
sec; the maximum specific resistance would then be about 101 3 Ωm, which is,
indeed, the order of magnitude mentioned in technical brochures for the prevention
of dust patterns on articles.

The tendency to form electrostatic charges could, therefore, be decreased by

increasing the conductivity of the polymer. To obtain drastic effects in this direction,
carbon black is added to the polymer; for less stringent requirements, antistatic
additives are used, which, blended into the polymer in small quantities, tend to
migrate to the surface and form a very thin surface layer, which considerably reduces
the surface resistance, and which is continuously renewed.

An even better effect is created by blending the polymer with metal fibres, e.g. for
shielding electric fields. Even a small amount of these fibres is able to form a
network, which considerably increases the conductivity.

A step further leads to the conducting polymers, a very special class of

macromolecules, which contain conjugated double bonds in the main chain
(alternating single and double bonds, -C=C-C=C-C=C-, such as polacetylene). As
such these polymers are not yet conductive; they are doped by adding or subtracting
electrons (e.g. with iodine or NH3 , resp.). The mobile charge carriers formed this
way provide the conductivity, which can reach values up to those of metals! An extra
advantage is that the oxydation and reduction processes are reversible. This opens
154 From polymers to plastics

the possibility to use these polymers e.g. as the positive electrode in rechargeable
batteries, together with, e.g., a lithium cathode.

8.2.2. Dielectric properties

The most important dielectric properties are the dielectric constant, ε , and the
dielectric loss factor, tan δ. These properties are of interest for alternating currents; ε
indicates the polarizability in an electric field, and, therefore, it governs the
magnitude of the alternating current transmitted through the material when used in a
capacitor. For most polymers ε is between 2 and 5, but it may reach values up to 10
for filled systems.

Of more importance is the loss factor, tan δ, denoting the fraction of the transmitted
alternating current lost by dissipation in the material. Here large differences occur
between polymers, as indicated in Figure 8.10. It appears that polymers with the
highest specific resistance also show the lowest dielectric losses. It should be
remarked, that the values given are very schematical; the losses are strongly
dependent on frequency and temperature.
tan δ

10 –1 PA, th.sets with fillers

10 –2

10 –3

10 –4

Figure 8.10. Dielectric loss factor.

As a matter of fact, these properties are highly relevant for electrotechnical

applications, in particular at high frequencies (losses are proportional to the product
of loss factor and frequency). The extremely low loss factor of polyethylene has
enabled the use of radar in the second world war.

But also in plastics processing technology the loss factor plays a role. A not too low
Further properties 155

value enables the heating-up of a polymer mass with the aid of a high-frequency
alternating electric field. The heat is then generated throughout the whole material,
so that the disadvantages of the low heat conductivity are evaded. Polymers with a
very low loss factor are not suitable for such a process: PE and PP films cannot be
welded by a high-frequency field. If such a process is explicitly required, for instance
to avoid shrinkage when welding biaxially oriented film, a thin layer of a second
polymer with a higher tan δ, is sometimes added by co-extrusion (see § 11.5.6).

8.2.3. Electric strength

For constructors of electrical equipment it is often of importance to know the
maximum electric tension which a material can bear without showing breakdown.
When studying electric failure, we notice an analogy with mechanical failure, which
was discussed in § 7.4: the electric breakdown tension is also time dependent. So the
question is, which level of stress can be applied so that within a certain period of
time no breakdown will occur. When we try to find this level by extrapolation from
short-term test results, we meet the same difficulties as with mechanical stresses: it
appears that failure can occur by two different mechanisms. At high tensions
breakdown is governed by the conductivity or the dielectric losses; the heat
developed brings about a temperature increase, often resulting in higher losses. This
goes on until the material softens or degrades, followed by electrical breakdown. At
lower tensions a narrow conducting path may develop slowly, often branched and
irregularly shaped, which eventually leads to local breakdown without overall heat
development (comparable to brittle crack formation, discussed in § 7.4.2).

The temperature plays, of course, a role in both mechanisms; the electric strength
decreases with increasing T. Moreover, though expressed in volts per meter
thickness, it depends on the thickness (again: non-linearity). Thin films have a
considerably higher strength than most values found in tables of properties, which
are results of tests on thicker samples.

As we noticed with the resistivity, the electric strength is also largely governed by
the nature of the surface. Contaminations may give rise to tracking currents along the
surface which locally develop so much heat that the material scorches and arc-
discharges occur, leading to fire risks. PVC and phenolic resins have a poor
resistance against tracking, as well as thermosets filled with wood flour. A high
resistance is shown by, i.a., PE, PMMA, PA and PS.

8.3. Optical properties

Most plastics transmit light, often better than glass. Transparency, however, is
156 From polymers to plastics

shown by fully homogeneous systems only. When more phases are present, the
components have, nearly always, a different refraction index, so that light is
scattered, resulting in opacity. This is the case with rubber-modified polymers (high-
impact PS, ABS etc.), with polymers containing fillers, but also with semi-crystalline
polymers. In the latter case crystalline and amorphous regions are present, which
differ in refraction index. Only if the crystallites are very small, and the scale of
dispersion is small with respect to the wave length of the light, can a semi-crystalline
polymer be transparent. This can be reached by quenching (e.g. thin PP film) or by
the application of nucleating agents. Also chain orientation, uniaxial as well as
biaxial, promotes the transparency; orientation creates a large number of
crystallization nuclei. By exception, the refractive indices of both phases may be
equal, such as in the transparent PMP.

Fully amorphous polymers may be transparent, such as PVC, PMMA, PC and PS.
They can, in principle, be applied in the optical industry for spectacles, simple
photographic lenses etc. For precision optics they are less suited, since because of
volume retardation as well as by the fact that they are often manufactured by
injection moulding, they cannot meet the requirements of narrow dimension
tolerances. Moreover, their low resistance to scratching is a disadvantage in optical

A special application of the high light transmittance of some polymers is the flexible
light conductor, the so-called fibre optics. This is a bundle of PMMA fibres, in which
each fibre is coated by a thin layer of another polymer, e.g. PE. Due to total
reflection at the wall, light can be transported without noticeable loss of intensity
along such a fibre, so that images can be transferred.

Though an amorphous glassy polymer may be perfectly transparent, it may show

optical anisotropy, which means that the refractive index depends on the direction;
the material is then birefringent. This birefringence is brought about by stresses such
as external loading or internal cooling stresses, but also by frozen-in rubberelastic
deformations or chain orientations. The latter are caused by the viscoelastic
behaviour of a molten polymer (see § 5.3.3), and are frozen-in when the material is
cooled to below its T g before the elastic stresses are relaxed. In some critical
applications, such as compact discs, birefringence is highly unwanted (see also §

8.4. Effects of environment

Properties of plastics are affected by the environment in various ways, as is also the
case with other materials, sometimes even to a much higher extent (wood, iron, etc.).
Further properties 157

Factors of importance are: oxygen in the air, light, water, chemicals, bacteria, etc.

Oxygen causes, at ambient temperature, a very slow oxydative degradation; at higher

temperatures the attack proceeds more rapidly. To prevent it, antioxydants are added,
depending on the type of application, which protect the polymer against oxidation in
a chemical way.

Light may also induce chemical breakdown; in particular the ultraviolet part of
sunlight is active in this respect. Here also additives are added for protection, the UV
stabilizers. Polymers which, by nature, are very sensitive to UV degradation, can in
this way be made excellently suitable for long-term outdoors application (e.g. PP). In
very demanding applications, the sunlight can be completely excluded by adding a
small amount of carbon black to the polymer. The resistance against sunlight can be
estimated on the basis of laboratory tests in which samples are exposed to a strong
light source with about the same spectral composition as sunlight. This enables a 3 to
5-fold acceleration of the effect. Sometimes also the weather is simulated in such a
test by alternating exposure to rain, radiation and temperature changes. Though these
tests provide a first impression of the weathering resistance, they cannot completely
replace tests in the real atmosphere. A series of results, obtained in different climates
and in the “Weather-o-meter” is, as an illustration, given in Figure 8.11.




Weather-o-meter Singapore
10 100 1000 days

Figure 8.11. Decrease of impact strength by ageing.

The consequences of oxydative and thermal breakdown of a polymer are:

discolouration, surface roughening, embrittlement, etc.; for rubbers: tackiness,
followed by embrittlement. For electrical applications oxidation goes accompanied
by a strong increase in the dielectric losses, and a decrease in insulation resistance
and breakdown strength.

Moisture has, in itself, usually not much effect on polymer properties, though the
amount of moisture which can be absorbed by polymers varies within wide limits
(between zero and a few %). Logically, the electric properties such as resistivity and
dielectric losses are the most sensitive to water. As to mechanical properties, nylons
show the strongest dependence on water absorption. PA-6 is able to take up a
158 From polymers to plastics

considerable amount of water, which acts as a plasticizer, in other words: it reduces

the glass-rubber transition temperature by some tens of °C.

Though in general polymers are non-corrosive and resistant against chemical attack,
they show, in some cases, weak spots. Ester groups are more or less sensitive to
bases and acids. A global impression of the chemical resistance of some polymers is
given in the following survey

Excellent resistance to acids as well as bases is shown by a.o. PTFE and PVC. Also
PS, PE and PP have a very good general resistance, but they are damaged by some
strong (oxidizing) acids in prolonged exposure. Polyamides, acrylates and cellulose
plastics are less resistant to some acids; sensitive for bases are, a.o., cellulose
plastics, polyamides, PC and some formaldehyde resins.

A different question is, whether polymers can withstand the influence of organic
liquids. Several of these are active as solvents or swelling agents, so that the
polymer, although not chemically attacked, loses its properties upon exposure. An
example is PVC, which softens and swells in acetone and in aromatic hydrocarbons
such as benzene. Most rubbers swell in aromatics, cellulose plastics are sensitive to
alcohol and acetone. PS is soluble in aromatics; practically all thermoplastics are
sensitive to chlorinated hydrocarbons, etc. A good resistance against most organic
solvents is shown by PTFE, PA, POM, PP and PE, the two latter ones, however, not
at elevated temperatures. Most thermosets are not affected by organic liquids.

8.5. Stress corrosion

Narrowly related to the effect of chemicals, discussed in the preceding section, is the
phenomenon of stress corrosion. This is the formation of cracks under the
simultaneous influence of a mechanical stress and a certain chemical environment;
neither the stress nor the environment can, separately, cause the same mechanical
damage, unless at a much longer time scale.

Typical examples of stress corrosion are: crack formation in strained rubber

vulcanizates under the influence of ozone, hair cracks in PE under stress in the
presence of a surface active agent (see also § 7.4.2), crack formation in PC, when
exposed to e.g. CCl4, within a few seconds after the application of a small stress.

The mechanisms of stress corrosion can be quite different: in the case first mentioned
the effect can be explained by a chemical reaction, which leads to irreversible
mechanical damage when the material is under stress. In the second case we could
think of a physical acceleration of a process of crack formation, which, otherwise,
would take a much longer time under stress; because of the lower interfacial tension
Further properties 159

less work is required to create new surface in the crack region. Since, in this case, no
chemical reactions take place, the expression “environmental stress-cracking” is to
be preferred.

General rules for the occurrence of stress corrosion cannot be given, since the
phenomenon is very specific for certain combinations of polymer and environment,
and is also dependent on the processing conditions in manufacturing the article (such
as the occurrence of cooling stresses and orientations). When plastics are used in
articles which are subjected to mechanical stress, such as pipes, crates, bottles,
screw-caps, etc., the risk of stress corrosion in the presence of fat, soap or organic
liquids, should always be taken into account. Proper choice of material and
dimensions can minimize the occurrence of crack formation.

8.6. Diffusion and permeability

Various low-molecular substances are more or less soluble in polymers, and are,
therefore, absorbed when in contact. Within the polymer such a substance can diffuse
to a place where the concentration is lower, and can leave the polymer elsewhere.
This means that the polymer is permeable to the low-molecular substance.
Obviously, the permeability, P, is governed by the solubility, S, and the rate of
diffusion, D; it appears that P = S·D.

The permeability of polymers to gases and water vapour is of practical relevance.

Between polymers considerable differences exist. For instance, for oxygen,
polyvinylidene chloride, PVDF, is 1000 times less permeable than low-density
polyethylene, LDPE. The permeability for liquids and vapours is strongly influenced
by the swelling which the polymer undergoes when taking up small quantities of
these substances. Even a minor degree of swelling brings about a large increase in
the rate of diffusion. Therefore, water vapour permeates much more rapidly than
gases; moreover the rate of permeation strongly depends on the water vapour

An impression of the order of magnitude of the permeability of some polymers for

nitrogen and for water vapour is given in the table below, in which the permeability
is expressed as the volume (cm3 ) passing per unit of area (cm2 ) when the layer
thickness is 1 cm, and the pressure difference of the gas or vapour on opposing sides
is 1 bar.
160 From polymers to plastics

for nitrogen:
0.01 ·10–10

for water vapour (95% relative humidity):

0.1 ·10–7
A few general trends can be recognized from this table, in particular as far as the
effect of diffusion is concerned. Rubbers, with their high free volume, allow a small
molecule to diffuse much faster than a glassy polymer. With semi-crystalline
polymers above Tg , such as PE and PP, the rubbery phase gives rise to a higher rate
of diffusion than when the amorphous phase is in the glassy state. Deviations from
this simple pattern can be attributed to differences in solubility (see also Qu. 8.16
and 8.17).

Plasticizers, antioxydants etc. sometimes show the tendency to migrate to the

surface, and thus to exude from the polymer. This tendency to migration is, in most
cases, unwanted, since the material gradually loses the properties obtained by the
presence of the additive, such as flexibility or chemical stability.

In some cases, however, the use of additives is based on their migration to the
surface: antistatic additives, and also agents to reduce friction of films, perform as
desired when they are present on the surface of the article, forming a thin skin.

A special problem is the migration of low-molecular substances in materials used for

food packaging. Here very stringent limitations on food contamination by
compounds which are possibly harmful for health, are necessary. This complicated
area is being studied carefully over the past decades, in order to establish for each
type of additive or monomer residue what its effect on the human organism could be,
and how much of such a substance could, under certain conditions, be expected to
migrate from the packaging into the food.

Polymeric compounds and

9.1. Polymer blends

9.1.1. General
Polymers are being blended in the fluid state. The first question is: are the polymers
miscible? We can distinguish between two cases:

- miscible; then a real blend is formed on a molecular scale; the individual

components have “disappeared”.
- unmiscible; we have a dispersion in which both components retain their own

Both types of blends are being used. Some examples:

- real blend: Noryl, a blend of polyphenylene ether (PPE) with polystyrene (PS).
PPE has a high T g , a high price, and is difficult to process. PS has a much lower
Tg , and is cheap. The blending ratio can be adjusted; three levels are
commercially available.
- dispersion: an example is high-impact polystyrene (TPS or HIPS), a dispersion of
rubber in PS. This blend is, necessarily, a dispersion, since impact energy should
be damped out in the rubber particles, which also stop developing cracks.
Analogous dispersions are high-impact modifications of PP, PA and PVC, and

In most cases polymer pairs are immiscible. Dispersions occur most often, real
mixtures are exceptions.

9.1.2. Miscibility of polymers

The first criterion of miscibility is of a thermodynamic nature:

GAB < GA + GB or ∆Gm < 0

(free enthalpy of mixing negative), so

∆Hm – T·∆Sm < 0.

162 From polymers to plastics

– ∆Hm is the enthalpy of mixing, and is governed by the interactions between the

∆H = V·(δ1 – δ2 )2 ·ϕ1·ϕ2 so> 0 .

δ1 and δ 2 are the solubility parameters, ϕ1 and ϕ 2 the volume fractions of the
components. With strong interactions (hydrogen bridges or polarity), H can be < 0
(exotherm); this promotes miscibility.

– ∆Sm is the gain in entropy upon mixing; it is always positive due to the greater
number of possibilities to place chain segments (more disorder in the mixture).
The entropy of mixing is strongly dependent on the chain length, or the degree of
polymerization, P.

Figure 9.1 presents a simple illustration for P = 3 in comparison with P = 1

(monomer); when the monomers (o and +) are blended, there are 20 possibilities for
their mutual positions in the blend (only one in the unblended condition), so:

∆S = k·ln (W2 /W1 ) = k·ln 20

for P = 3 there are only six possibilities, so ∆S = k·ln 6.

unblended blended



Figure 9.1. Simple illustration of entropy of mixing.

Application of conformation statistics results in a value of ∆Sm. Combination with

∆Hm gives the Flory-Huggins relation:

∆Gm ϕ1 ϕ2 χ
= kT· [ V1
·ln ϕ 1 +
·ln ϕ 2 +
V s · ϕ 1 ·ϕ 2 ]
Polymeric compounds and composites 163

where ϕ1 and ϕ2 are the volume fractions, V1 and V2 the sizes of the molecules, χ the
interaction parameter and Vs the interaction segment size. It appears that the first two
terms, the entropy contributions, both negative, approach to zero with strongly
increasing chain length.

The first condition for miscibility, ∆G m < 0,, is not always sufficient for the
formation of a homogeneous blend. The kinetics of blending also plays a role; in the
blending process the domain sizes have to be sufficiently reduced before complete
mixing by diffusion can take place. Diffusion is a slow process, so that homogeneous
mixing can only be reached after a long time of mixing.

Moreover, not only the sign of ∆Gm counts, but also the shape of the curve of ∆Gm
versus volume fraction ϕ . The simplest case is shown in Figure 9.2a, where the sign
of ∆Gm is the decisive factor. The situation is more complicated with, for instance,
Figure 9.2b; though ∆Gm is here negative for all blending ratios, the maximum in the
curve at about ϕ = 0.5 causes an instability.

Independent of the level of the curve (of the sign of ∆G m ), a small disturbance,
caused by a local difference in concentration, will bring about an increase of ∆Gm)
when ∂2 Gm/∂ϕ12 > 0, and will, therefore, disappear; the blend is stable (Figure 9.2c).
When the curve is, however, convex, with ∂2 Gm/∂ϕ12 < 0,, a minor distortion will
grow, initiating segregation; the system is not stable (Figure 9.2d).

For a situation where ∂2 Gm/∂ϕ12 > 0, but in the vicinity of a point of inflection
(∂2 Gm/∂ϕ12 = 0), a large disturbance may initiate segregation; the system is in a
metastable condition (compare an upright standing pencil). The point of inflection is
called the spinodal (Figure 9.2b and 9.2e).

∆G m incompatible, ∆G m > 0 spinodal

∆G m

compatible ∆G m < 0
a. b.


∂2 G m > 0 ∂2 G m < 0
stable unstable
∂ϕ12 ∂ϕ12 metastable
c. d. e.
Figure 9.2. Various cases of miscibility.
164 From polymers to plastics

With changing temperature the curves change; for spinodals and binodals (the points
of contact with the common tangent), plotted as T against ϕ1, a picture as shown in
Figure 9.3 may be obtained. In this case a “lower critical solution temperature”,
LCST, exists, which is a maximum temperature at which, for every blending ratio, a
stable homogeneous blend is possible. An “upper critical solution temperature”
(UCST) may also exist; in that case the curves are upside down. Besides, more
complicated cases are possible (see also Qu. 9.4 tu. 9.7).
stable stab.
∆G m



T spinodal


lower critical
solution temp.
0 ϕ1 T

Figure 9.3. Binodals and spinodals.

9.1.3. Detection of miscibility

Microscopic or electron microscopic observations on solid blends provide
information on the homogeneity or the degree of dispersion in a polymer blend. An
effective method is also the analysis of the glass-rubber transition(s). The presence
of a single glass-rubber transition, in between the transition temperatures of the
components, indicates homogeneity of the blend (see also § 3.5). However, the glass
rubber transition occurs when large chain parts, of about 20 to 60 segments, acquire
free mobility. A single Tg, therefore, indicates homogeneity on that scale, and not
necessarily on the scale of single segments.

Tg can be determined by dynamic mechanical experiments from the log E–T

diagram, but also from the maximum in the tan δ - T curve. Another possibility is
differential scanning calorimetry (DSC).

Figure 9.4 gives some possible pictures of log E-T curves (already partially
Polymeric compounds and composites 165

mentioned in § 3.5). Figure 9.4a refers to one extreme: a homogeneous blend, where
the curve does not show any sign of the transition of the individual components.
Figure 9.4b holds for the other extreme: it denotes the behaviour of a pure
dispersion, in which A and B retain their own individuality.

Figure 9.4c and 9.4d represent intermediate cases; 9.4c indicates partial miscibility;
we see a two-phase system of AB blends with different A/B ratios. This might be the
result of segregation into the binodals. Figure 9.4d is called an “interphase” or a
“multiphase” blend. The system is quasi-homogeneous, but it contains all A/B ratios
between ϕ1 = 0 and ϕ 1 = 1. It looks like a system with concentration gradients as a
result of non-completed diffusion in a combination of well-compatible polymers.

log E log E


a. T b. T

log E log E


c. T d. T

Figure 9.4. E(T) for various types of blends.

9.1.4. Block copolymers

Block copolymers are a special case of two-phase systems. They are not real blends,
since their components are tied together in one and the same chain. An example is:


If the homopolymers A and B are incompatible, the copolymer may show

segregation. The A-segments and the B-segments are then located in separate
domains (Figure 9.5).

The following questions are relevant:

- when does segregation occur?

166 From polymers to plastics

- what is the size of the domains?

- how are the domains shaped?

Figure 9.5. Domain formation in a block copolymer.

In some cases these questions can be answered by rigorous thermodynamical

treatment, again on the basis of the formula ∆G = ∆H – T·∆S, in which ∆ now
denotes the difference between between the randomly mixed and the segregated
condition. The approach starts, compared with that on mixing, at the opposite side,
namely not from the conformation possibilities in the mixed condition, but now in
the segregated system.

The entropy difference is written as:

∆S = ∆S1 + ∆S2 + ∆S3

in which each of the terms represents one aspect of the ordering into domains,

∆S1: all A-segments are present in an A-domain and all B-segments in a B-domain;
∆S2: All A-B connections are situated in the interface between A and B-domains;
∆S3: Within a domain the chain conformations are restricted in comparison with a
“random flight” situation.

Moreover, an interfacial free energy term, Gs, is present, so that:

Polymeric compounds and composites 167

∆G = ∆H + Gs – T(∆S1 + ∆S2 + ∆S3 )

The result of this approach for the three-block copolymer SBS (styrene-butadiene-
styrene) is as follows:

- The critical molar mass, above which segregation occurs, Mcr, is 2.5 to 5 times
greater then for equivalent blends. This means a better “miscibility” of block
copolymers, as a result of the entropy restrictions in the domain structure.

- The domain size can be calculated; for a certain copolymer it amounts, e.g., to 30

- The morphology depends on the A/B ratio: when A/B < 0.2, ∆G is smallest for
spheres, up to 0.3 for cylinders , and above 0.3 for lamellae. At A/B ≈ 0.5 we find
lamellae of A and B; at higher A/B the pattern repeats itself in the reverse
direction, while the roles of A and B are exchanged (Figure 9.6). With cylinder
morphology the cylinders can be oriented in elongational flow; by further
annealing they grow into continuous “fibre reinforcement”. The properties are
then strongly anisotropic: the stiffness in the direction of orientation is many
times higher than across (if the cylinders are polystyrene domains). After special
annealing treatment highly regular structures have been obtained with cylinders
(with a diameter of 20 to 30 nm) packed in a perfect hexagonal order. But also an
extruded sheet clearly exhibits anisotropy.

∆G sphere cylinder

0,1 0,2 0,3 0,5 M A /M B

sphere cylinderlamella

Figure 9.6. The morphology depends on the monomer ratio.

The desirability of segregation in block copolymers can be demonstrated by

considering the behaviour of SBS, which is one of the oldest types. It has about the
same chain composition as SBR, but, rather than SBR, it shows two glass-rubber
transitions, namely that of polybutadiene and that of polystyrene. Between these two
temperatures it behaves as a rubber, in which the PS domains act as cross-links; it is,
therefore, a “self-vulcanizing” rubber (see also Figure 3.8; see Qu. 9.14). Moreover,
the hard domains play the role of a reinforcing filler.

Above the Tg of PS, the domains soften, so that the polymer can be processed as a
168 From polymers to plastics

thermoplast, a “thermoplastic rubber” or “thermoplastic elastomer, TPE” . In the

fluid condition polybutadiene and polystyrene are still incompatible, so that domains
are continuously being formed and disrupted. The viscosity is, therefore, higher than
that of SBR with the same S/B ratio. Moreover, the domains in the fluid are
responsible for the existence of a small yield stress.

Another example is even older: the thermoplastic elastomers based on polyurethane;

their chains contain a number of blocks, alternatingly “hard” and “soft” PU blocks.
(elastic yarns, such as “Lycra” and “Spandex”).

If the hard blocks are longer than the soft ones, such as in SBS with a high styrene
content, the hard phase will be continuous, and the rubbery phase is present as
domains (see Figure 9.6). In such a case SBS behaves as a high-impact PS. Another
example of this type is a PP/EP block copolymer; tails of EP (random copolymer of
ethylene and propylene) on the PP chains segregate into rubbery domains in the PP
matrix, which improve the impact strength.

Newer developments are: polyester/polyether block copolymers (Hytrel, Arnitel),

etc. By choosing various levels of block length ratio, a broad spectrum of stiffnesses
(or hardnesses) can be obtained, which practically fills the gap between rubbers and
thermoplasts. TPE’s form a rapidly growing class of materials, which find an
increasing number of applications.

9.1.5. The formation of dispersions

When immiscible polymers are blended in the fluid condition, a dispersion is
formed. Starting from the original dimensions (granules or powder), the size is
gradually reduced down to m scale. In the mixing device the melt is subjected to
shear flow and elongational flow; in both of these the domains deform and can break
up. When a sphere is deformed into an ellipsoid, its surface area increases and thus
its interfacial energy, until the retracting force, exerted by the interface, balances the
force exerted by the flow field. Larger droplets hardly meet any counteracting force;
the smaller ones attain a stable deformation.

The governing parameter in this process is the Weber number, Ca = 1/k, with

σ σ/R σ/R
k= = =
γηcR γ ηc τ

in which σ = interfacial stress

γ = rate of shear
ηc = viscosity of the continuous phase,
R = initial radius of the sphere,
Polymeric compounds and composites 169

τ = shear stress


Figure 9.7. Deformation of a spherical droplet.

Cox deived:

L–B 5(19λ + 16)

D =L + B=
4(λ + 1)[(19λ )2 + (20k)2 ]1/2
viscosity droplet
λ = ηd /η c = viscosity continuous phase

For λ < 1 this can be approximated by:

1 19λ + 16 1 γηcR
D ≈k · ≈ =
16λ + 16 k σ

The droplet breaks at D ≈ 0.5, dus γηcR ≈ σ /2. This is Taylor’s formula for the
critical shear rate at which break occurs:

σ 16λ + 16
γb = ·
2ηcR 19λ + 16

Smaller droplets need a higher shear rate to break. At constant γ there is an

equilibrium drop size:


Complications may arise as a result of:

– non-Newtonian behaviour ( η depends on γ),

– elastic behaviour of molten polymers.

It is possible to take these effects into account, but the actually found scale of
dispersion in practical blending operations is never properly matched by these
calculations; the observed particles are much bigger (up to ten times) than according
to the Taylor equation. The explanation of this discrepancy is coalescence: small
droplets join together when they collide.
170 From polymers to plastics

In a collision, droplets are flattened; a condition for the occurrence of coalescence is,
that, within the time scale of passing each other, the layer of liquid polymer between
the droplet can be squeezed out until a critical distance of approach has been
reached. The rate of approach of the flattened surfaces strongly depends on the
nature of the interface; this can be “rigid” or “mobile” :

- rigid: the interface can withstand tangential stresses without flow occurring
along it.
- mobile: the interface is able to move.

rigid mobile

v v

Figure 9.8. A rigid and a mobile interface.

The velocity profile in both cases is sketched in Figure 9.8; with a mobile interface
the rate of approach is much higher than with a rigid one, resulting in a higher
probability of coalescence. In most cases a rigid interface is met, as a result of
contaminations; very low concentrations of a surface active agent (10–10 g/mole) are
already sufficient. With polymer-polymer interfaces there are, however, indications
that the interface is mobile. This means that, in a blending process, a continuous
succession of break-up and coalescence occurs, so that the resulting droplet size is
much greater than expected from Taylor’s formula. It has been observed that at very
low concentrations (< 0.5 %) small droplets, approximately according to Taylor, are
formed; the chance of collision is then much smaller. With increasing concentration
the droplet size increases strongly.

The role of surface-active agents may, with a pair of polymers A and B, also be
played by a two-block copolymer AB. It has been observed that a small addition of
AB results in a considerably finer dispersion in the blend. This can be explained in
two ways:

– The interfacial tension is decreased as a result of the presence of AB; the Taylor
formula then predicts a smaller droplet size;
– The interface becomes rigid, so the chance of coalescence decreases.

If the interfacial tension is too small or the droplet too big, the retractive forces may
be too small to counteract the droplet deformation; in such a case threads may be
formed. Such a thread is, in general, not stable; a small distortion as a deviation from
Polymeric compounds and composites 171

the cylindrical shape may grow (Rayleigh, 1878). The growth of such a distortion is
schematically represented in Figure 9.9.

Figure 9.9. Instability of a cylindrical fluid thread.

— —
When R(z) = R + α·sin(2πz/ Λ ) the thread is unstable for Λ > 2π R; with increasing
amplitude α the total surface area will then decrease (which is, at first sight,
surprising!). α then grows exponentially:

α = α0 ·exp qt

q is a complicated function of all variables; it shows a maximum as a function of the

wave length Λ at Λm. Λm is, therefore, the “dominant wave length”: with an arbitrary
pattern of disturbances (a broad spectrum of wave lengths) the disturbance with Λm
will grow the fastest and will lead to a split-up of the thread into droplets.

This splitting-up is driven by the interfacial tension, which causes pressure

differences between parts of the thread with positive and negative curvature. Only
the viscosity counteracts the growth of the distortion and limits its rate of growth.

Fluids with a small yield stress may, however, be stable, namely if this yield stress
exceeds the pressure differences. An example: the three-block copolymer SBS
exhibits a minor yield stress, which is hardly detectable, but sufficient to prevent the
growth of distortions. As discussed in § 9.1.4, it is caused by the polystyrene
domains. It has also been observed that, when blending SBS with another polymer, it
shows, even at very low concentrations, a pronounced tendency to form a continuous
phase, in other words: it does not break up into dispersed particles.

If the break-up time of threads is longer than the time of cooling, threads will also
remain present after solidification. These are, however, not permanent; upon further
processing by extrusion or injection moulding, the instabilities can grow further.

9.1.6 Properties of blends

An important question is, what the properties are of a dispersion blend of two
polymers, A and B, when the properties of the components are EA and EB, and the
volume fractions ϕA and ϕB. First we consider the two simplest cases: a series and a
parallel arrangement.
172 From polymers to plastics


series parallel

Figuur 9.10 Series- and parallel arrangement.

For the parallel case, E (e.g. the modulus of elasticity, the electric or the thermal
conductivity), can be expressed by :

E = ϕA·EA + ϕB·EB

and for series:

1 ϕA ϕB
E = EA + EB

For, e.g., the electric resistance and the compliance the same formulae hold, but

The general way to express the various quantities is:

Main phase: volume fraction 1 – ϕ

property E0

Additional phase: volume fraction ϕ

property E1
E1 E
E0 = α E0
= f(α,ϕ)

Parallel: E = 1 + ϕ(α – 1)

E 1
Serie: E =
0 1 – ϕ(α – 1)/α

Combinations of series and parallel provide a more realistic representation of the

properties of a dispersion; instead of ϕ then, however, more parameters are required,
e.g. (Takayanagi):

Further splitting-up necessitates the use of an increasing number of parameters, while

the equations become more and more complicated.

A better theory for dispersed spheres has been given by Halpin and Tsai and by
Polymeric compounds and composites 173

1–λ E0
λ E1 E0

µ 1–µ

Figuur 9.11 Takayanagi model.

α–1 E1
1 + ϕA with α = E
E α+A 0
E0 α–1 7 – 5ν
1–ϕ and A =
α+A 8 – 10ν

This formula is valid for the modulus of elasticity, while ν is the Poisson ratio of the
matrix material. It is only valid up to ϕ ≈ 0,7 (near the tightest packing of spheres).
For higher values, from ϕ ≈ 1 going down to ϕ ≈ 0,3 (or lower) the structure can be
reversed: the roles of the components are exchanged. We can use the same formula,
but now with 1 – ϕ instead of ϕ, while α is replaced by 1/α. For α = E1/E0 = 100, the
curves, calculated from both formulae, are given in Figure 9.12.
Before the tightest stacking of spheres has been reached, the formula is subject to
deviations, in particular if the embedded particles are not deformable, such as is the
case with hard particles. The denominator of the formula is then modified as follows:

1 – ϕ·ψ·(α – 1)/(α + A)
in which
ψ = 1 + ϕ·(1 – ϕm)/ϕm2.

ϕm is the volume fraction at the tightest packing of the dispersed phase, which for
equal spheres is about 0.74, so that then ψ = 1 – 0.475·ϕ. The curve is then steeper,
as indicated in Figure 9.12. Up to ϕ = 0.4 the difference is neglegible. With spheres
of differrent sizes a tighter packing can be reached; the deviations are then smaller.
With the inverse morphology (soft phase dispersed in a hard matrix) a similar effect
occurs; the effect is, however, less pronounced, because soft spheres can more easily
deform and can, therefore, attain a higher degree of filling (this is, in particular, true
for foams).

At some volume fraction, mostly around ϕ = 0.5, phase inversion occurs; around this
value, A and B are both continuous. This may, however, also happen at other values
of ϕ , depending on the rheological behaviour of the components. Moreover, co-
continuity is possible at any volume fraction (see in § 9.1.4 the remark on SBS). For
the properties of a co-continuous two-phase system an expression has been derived
by Nielsen:
174 From polymers to plastics

ϕ (α = 0.01)
1 0.8 0.6 0.4 0.2 0


Kerner Nielsen
30 parallel A in B n = 1/5

with ϕm


5 B in A


0 0.2 0.4 0.6 0.8 1

Figuur 9.12 Stiffness as a function of composition.

En = (1 – ϕ)·E0n + ϕ·E1n

Extreme cases of this equation are:

n = 1 parallel arrangement
n = –1 arrangement in series
n = 0 log E = (1 – ϕ)log E0 + ϕ log E (straight line on a log-log scale)
1 1
Mostly n = 3 for heat conduction, n = 5 for the modulus of elasticity.

Figure 9.12 shows the various possibilities. The figure demonstrates the importance
of the type of dispersion with respect to the properties. When B is a stiff polymer (E
= 100), and A a soft one (E = 1), we read at ϕB = 0.25 the following values for E:
Polymeric compounds and composites 175

A dispersed in B: E = 16
B dispersed in A: E = 1.8
A and B co-continuous: E = 5

As another example we consider two (amorphous) polymers with strongly different

Tg , namely high for A and low for B. If we choose ϕ B = 0.4, then for each of the
three types of dispersion the E -modulus can be calculated as a function of
temperature from the E(T) curves of the components. The result is shown in Figure
9.13, in which also the values of the ISO and Vicat softening points have been
schematically indicated. It again appears that the type of dispersion has a strong
effect on the properties (see also Qu. 4.24, 9.13, 9.15 and 9.23).
log E ϕB = 0.4

A in B


A B in A

Figuur 9.13 Effect of morphology on softening temperatures.

Morphology control in a mixing operation is, therefore, very important for utilizing
the properties of a blended-in high-performance polymer. It appears that continuity
of one or both phases is often essential to obtain optimum properties. This continuity
is generated in the blending process, though, unfortunately, not much is known of the
ways to control it.

Continuity can also be reached by polymerizing one of the components within the
other. In such a case the blend is called an IPN, an interpenetrating network; it is, in
most cases formed by a thermoset in a thermoplastic polymer. An example is a
compound built-up from 50% of a thermoplast (polycarbonate or polysulphone), and
50% of a cross-linked polymer on the basis of dicyanate bisphenol-A. The skeleton
176 From polymers to plastics

of the thermoset in the thermoplastic matrix provides the material with exceptional
properties. It combines the toughness and the strength of the thermoplast with the
high resistance to temperature of the thermoset. When PC is used as the matrix a heat
deflection temperature of 200 °C is realized (for PC: 135 °C).

9.2. Reinforcement by particles

Very often particles are blended into polymers, in thermoplasts as well as in
thermosets and in synthetic rubbers. This is done for various reasons: the aim may be
stiffness, strength, hardness, softening temperature, a reduction of shrinkage in
processing, reduction of thermal expansion or electric resistance, or, simply, to
reduce the price of the material. The fillers used are: wood flour, carbon black, glass
powder, chalk, quartz powder, mica, molybdene sulphide, various metal oxides, etc.

An old example of a polymer reinforced by particles is rubber with carbon black.

Before 1920 this was unknown; with other, less effective fillers an automotive tyre
was worn-out after 8,000 km. The much longer life (100,000 km or more) of modern
tyres is, of course, due to a number of factors, but is largely brought about by the
enormous effect of carbon black on the abrasion resistance. The reinforcing effect of
carbon black is also demonstrated by its effect on the tensile strength of rubber
vulcanizates. An unfilled, vulcanized SBR breaks at a load of 2 to 3 MPa; with
carbon black the tensile strength exceeds 30 MPa. Such an enormous reinforcement
is not met with any other polymer, neither with particles, nor with short fibres.
Apparently something special is happening. In the first place the enormous surface
area should be mentioned. Carbon black forms agglomerates, which consist of
aggregates of the primary particles of a few tens of nm. In a usual formulation, for
instance, 50 grammes of black are added to 100 grammes of rubber (the volume
fraction of black is then 0.2), so that in a small piece of rubber thousands of m2 of
carbon black surface are present. Along this surface the polymer chains are tightly
adsorbed, though they are able to slide along the surface in the direction of the chain;
this results in a levelling of the stresses in the chain parts and thus an optimum

In most cases the main effect of particles is an increase in stiffness, up to two or three
times its original value. In the same way as with polymer blends, this increase in
stiffness can be calculated theoretically, namely with the same Kerner formula as in
§ 9.1.6.
Polymeric compounds and composites 177

1 + ϕA
E α+A
E0 α–1

It is interesting to consider an extreme case for very stiff spheres (α very high) at a
low volume fraction. The Kerner equation can then be written as:

E/Ec = 1 + ϕ(A + 1) = 1 + kE·ϕ

With rubber as a matrix (ν = 0.5), k E has the value of 2.5, and we meet the well-
known Einstein formula for the increase in viscosity as a result of the presence of
spherical particles.

For non-spherical particles the relations are different; this is easiest demonstrated by
the Einstein formula: the coefficient kE increases when the particles are flattened, and
even stronger upon elongation. An example of the latter case: When a sphere is
extended up to a length/diameter ratio of 10, kE increases from 2.5 to 6. Sticks and
platelets, therefore, have a stronger effect; this accounts for the fact that with a
volume fraction of 20 % increases in stiffness up to a factor of 3 are measured,
whereas the Kerner formula would predict a much lower value.

In contrast to the stiffness, the strength of polymers is hardly influenced by

“reinforcing” fillers. In most cases we observe a reduction in the ultimate strain, and
often a small decrease in tensile strength and in impact strength. Here the adhesion
between polymer and particle plays a substantial role; when a high shear stress
induces dewetting at the interface, the chance of crack initiation is high. But also
with good adhesion the particles form regions of stress concentration, which may
initiate crack formation. There are exceptions in which the strength increases; the
most drastic one is the effect of carbon black in rubbers, mentioned before.

9.3. Short fibres

In the foregoing we have seen the influence of the shape of particles om the increase
in stiffness: longer particles have a greater effect. A logical consequence is the use of
fibres. Most thermoplastics, in particular when they are intended for use as
construction materials, are commercially available in formulations with short glass
fibres. These formulations are manufactured as nibs, made in the usual extruder and
a granulator, after a blending process in which the fibres are blended into the molten
polymer. Besides glass fibres, also carbon fibres are being applied. The fibres are a
few mm long, and, e.g. 10 µm thick. During blending and further processing fibres
178 From polymers to plastics

may break, so that their final length may differ from case to case.

To get some understanding of the effect of short fibres on the properties of the
composite material, we can start off with considering a single fibre, embedded in the
polymer in the direction of the tensile stress (Figure 9.14). The fibre has a much
higher stiffness than the polymer, but will only be able to contribute to stiffness and
strength of the composite if it carries part of the load. At its extremes this is not the
case; the cylindrical interface is subjected to shear loading, which, with a good
adhesion, gradually decreases from the fibre end, while the tensile load in the fibre
increases. The effective part of the fibre is, therefore, its total length diminished by
the length of the transient parts. The fibre should, therefore, not be too short, so that
its load-bearing function would be hampered.

tensile stress

shear stress

Figure 9.14 Stress distribution round a short fibre.

For the quantitative description of the stiffness of the composite, the Kerner equation
of § 9.1.5. and § 9.2 can again be applied, though in a somewhat modified shape.
Since a fibre brings about anisotropy of properties in its environment, we have to
consider two different cases, viz. the E-modulus parallel to the fibre, Ep , and the one
perpendicular to it, Et. For both cases Kerner’s equation holds, in which now A, in
the parallel case, Ep , equals 2·l/d, while for Et, A = 2 .

A convenient approximation is found by considering a small value of ϕ; then:

E α–1
= 1 + ϕ · ·(A + 1)
E0 α+A

It follows that for long fibres (A >> α):

E0 = 1 + ϕ(α – 1)

or, with E1 = α·E0 as the modulus of the fibre:

Polymeric compounds and composites 179

E = E0·(1 – ϕ) + E1·ϕ

which, as could be expected, turns out to be a simple parallel arrangement of both

components. This also appears from Figure 9.15, in which, for two values of α, the
relative moduli of the composite, for ϕ = 0.2, are plotted as a function of l/d (upper
two curves). For α = 25, a maximum value which is representative for glass fibres in
a hard polymer, of 0.8 + 25 · 0.2 = 5.8 is about reached at l/d = 200. For α = 125, an
average value for carbon fibres in a hard polymer, the maximum of 0.8 + 125 · 0.2 is
not nearly reached at the same fibre length; with better, stiffer and more expensive
fibres it is, therefore, the more important to prevent rupture of fibres during
processing. It also appears that for very short fibres the nature of the fibres (cheap
glass fibres or expensive carbon fibres) is of less importance.



3 parallel
1 length/diameter
1 10 100 1000

Figure 9.15 Stiffening as a function of fibre length and orientation.

These values hold for the case that the fibres are oriented in the stress direction. In
reality the orientation of the fibres is at random in three dimensions; we should,
therefore, not only consider Ep but also the much smaller Et, which, even for very
stiff fibres, is not higher than about E0(1 + 1,5·ϕ ) or, with ϕ = 0.2, 1.3·E 0 A
frequently used expression for the modulus of an isotropic composite with
chaotically arranged fibres, is:

E = 0.2 · Ep + 0.8 · Et

For both values of α this approximation is also plotted in Figure 9.14; the effect of
fibres on the stiffness is then strongly “diluted”, but is the same in three dimensions.

We can now compare these calculated values with test results. Technical brochures
issued by raw material manufacturers supply an abundance of data on their products,
also on the tensile moduli of the virgin polymers and on those containing a certain
weight fraction of glass fibres. Combination of these data for a number of polymers
(POM, PPE/PS, PC, PBTP, PET, PA-6) results in Figure 9.16. We can read from this
180 From polymers to plastics

figure that the average stiffness ratio at 20 vol % glass is about 4. The theoretical
curve for isotropic composites in Figure 9.14 yields a value of only 2.2, and for
completely oriented fibres a value of 5.8. From a different theoretical approach, not
to be discussed further, a ratio of only 1.9 is obtained for the isotropic case.
Apparently the experimental data have been obtained on test bars in which, during
their formation, the fibres were considerably oriented. These data are, therefore, not
representative for the properties of an arbitrary end-product.

2 chaotic

vol% glass fibres

10 20 30 40

Figure 9.16 Increase in tensile modulus as a result of the presence of short glass fibres.

Contrary to particulate-filled composites, fibres induce an increase in strength. Again

on the basis of data from technical brochures, it can be concluded that the tensile
strength is raised by a factor between 1.5 and 3.5 at 20 vol % glass fibres. Apparently
the fibres play such an important load-carrying role, that the polymer itself is
considerably less stressed. The strain at break, however, strongly decreases.

The effect of short fibres on the impact strength varies; sometimes it is positive and
sometimes negative. Broadly speaking, polymers with a low impact strength, and
also tougher polymers at low temperatures, are improved in this respect. Also the
reverse holds true: tough polymers undergo a loss in impact strength when they are
filled with glass fibres. Figure 9.17 shows schematically the region in which the
impact strength values for a number of polymers, containing different amounts of
glass fibres, are found. Fibres appear to level out the differences in impact strength of
the various polymers.

The effect of short fibres on the softening temperature has already been discussed in
§ 8.2.2. It appeared that their effect on amorphous polymers is minor, whereas for
semi-crystalline polymers the softening points are drastically increased. Both
observations are directly related to the slope of the log E - T curve.

Besides for improving mechanical properties, short fibres can also be applied for
other purposes, such as electric conductivity. Metal fibres, if dispersed in such a way
Polymeric compounds and composites 181

that they touch each other, are suitable in this respect. Though the conductivity is not
high enough to enable transport of energy, it is adequate for electromagnetic




vol% glass fibres

10 20 30 40

Figure 9.17 Impact strength as a function of glass fibre content.

9.4. Long fibres

For making composites of a polymer with long, continuous fibres, quite different
techniques are used than for short fibres. Now it is no longer the polymer which,
filled with reinforcing fibres, undergoes a formation process to produce a finished
article. Now the fibre mass is formed and, thereafter, filled with the polymer. The
positions are inverted, not only in the manufacturing of the composite, but also in its
properties, since now its behaviour is largely governed by the fibres rather than by
the polymer.

A first consequence is, that as matrices for long fibres thermosets are primarily
suited. Fibre bundles or cloths can be much more easily impregnated with a low-
molecular, low-viscosity resin than with highly viscous molten thermoplastics. Most
of the long-fibre composites are, therefore, based on thermosets, mainly polyesters
(UP) or epoxies (EP). However, during the past few years thermoplastic matrices are
being used at an increasing scale; impregnating is carried out with special moulding
techniques (see § 11.6.2).

Glass fibres are being used the most frequently; besides, for demanding applications,
also modern fibres such as carbon and aramide, sometimes in combination with each
other. The fibres can be arranged in rovings or in cloth, built-up from rovings or
from twisted yarns. With rovings it is possible to achieve unidirectional orientation;
with cloth the effect of the reinforcement is in two directions. Felt-liied, resulting in
a three-dimensional reinforcement, though, as a matter of fact, to a lower degree than
182 From polymers to plastics

with the anisotropic structures.

An illustration of the effect of the distribution of fibre directions is given in fig. 9.18.
For glass fibre reinforcement in polyester resin some average values of tensile
modulus and strength are plotted, all for 40 weight % (25 vol%) glass fibre.
Moreover the properties of the pure resin and of the glass fibre are indicated. For the
case of parallel fibres the test results are in good agreement with the simple rule of
additivity mentioned before; also in the two other cases a reasonable agreement with
theory is found.

MPa tensile strength

glass fibre
UP with parallel
500 40% w.
glass fibres cloth
square weave

UP resin
1 2 5 10 20 50 100 GPa

Figure 9.18 Effect of orientation distribution on properties.

Figures 9.19 and 9.20 present a survey of the mechanical properties of some
(unidirectional) composites, in comparison with some other materials. In Figure 9.19
the values of modulus and strength are plotted as such, while in Figure 9.20 these
values have been divided by the specific mass. From Figure 9.20 the enormous
advantage of composites with respect to stiffness and strength per unit weight, in
comparison to metals, is clearly visible. The modern carbon and aramide composites
are superior to those based on glass fibres, for the specific stiffness even by a factor
between 4 and 5.

The data in Figure 9.20 are of particular importance in applications where inertia
effects and/or gravity forces should be small, e.g. in ski’s, tennis rackets, motorcars,
aeroplanes, space vehicles etc.
Polymeric compounds and composites 183

10.000 tensile strength
composites steel


10 E-modulus
1 10 100 1000 GPa

Figure 9.19. Mechanical properties for some composites.

tensile strength/density

glass aramide

100 steel


10 E-modulus/density
1 10 100 1000 GPa/(Mg/m3)

Figuur 9.20. Mechanical properties for some composites, divided by density


Data on materials
Several properties of polymers have been dealt with schematically in the foregoing

Extensive tables of data on materials would be beyond the scope of this book; they
can be found abundantly in handbooks and in the many technical brochures, issued
by manufacturers.

Therefore, a few brief surveys, by no means complete, of some important properties

are given on the following pages.

When using these data, it should be kept in mind that their value is, in most cases,
very limited. A first reason is the strong dependence of all properties on temperature
and on time, as discussed before, so that the data given can certainly not be
considered as material constants.

In the second place, within a given type of polymer a broad variety of grades exists,
as regards small variations in chain structure, differences in chain length and its
distribution, and the presence of various additives for improving impact strength, fire
resistance, stability etc.

Finally, it should be realized that, even with a well-defined material in a standard test
method, the properties may be dependent on the way in which the test sample has
been made; the preparation conditions may be different, and, moreover, may
strongly deviate from those present in a technical processing operation. This applies,
e.g. to shrinkage stresses, nature of the surface, crystalline structure, and, in
particular, chain orientations.
Data on materials 185

10.1. Thermoplastics
Abbre- Name Density Price-
viation kg/dm3 class*
LDPE Low density polyethene 0.92 1
HDPE Highdensity polyethylene 0.95 1
PP Polypropylene 0.90 1
PVC Polyvinylchloride 1.38 1
PS Polystyrene 1.05 1
TPS High impact polystyrene 1.05 1
SAN Styrene-acrylonitril copolymer 1.08 2
ABS Acrylonitrile-butadiene-styrene copolymer 1.05 2
ASA Acrylonitrile-styrene-acrylate copolymer 1.07 3
PMMA Polymethylmethacrylate 1.18 2
PA 6 Polyamide-6 (nylon-6) 1.13 4
PA 6.6 Polyamide-6.6 (nylon-6.6) 1.13 4
PA 11 Polyamide-11 (nylon-11) 1.04 4
PA 12 Polyamide-12 (nylon-12) 1.02 4
POM Polyoxymethylene 1.42 3
PC Polycarbonate 1.22 4
PETP Polyethylene terephtalate 1.37 1
PBTP Polybutylene terephtalate 1.29 3
PPE/PS Polyphenyleneether + polystyrene 1.06 3
PSU Polysulfone 1.24 5
PPS Polyphenylenesulfide 1.34 4
PI Polyimide 1.43 6
PTFE Polytetrafluorethylene 2.17 6
FEP Hexafluorpropylene-tetrafluorethylene copolymer 2.15 7
PVDF Polyvinylidenefluoride 1.78 6
ETFE Tetrafluorethylene-ethylene copolymer 1.70 7
CA Celluloseacetate 1.30 3
CAB Celluloseacetate-butyrate 1.20 3
PB Polybutylene 0.92 2
PMP Polymethylpentene 0.83 4
PEEK Polyether-ether-ketone 1.30 7
PES Polyethersulphone 1.37 7
PK Polyketone 1.24 3

* price class 1 2 3 4 5 6 7
€/kg 0.5–1 1–3 3–5 5–8 8–14 14–25 25–55
186 From polymers to plastics

Abbrevia– E-modulus Tensile Strain at Notched

tion (short-term) strength fracture impact
(short-term) strength
MPa MPa % kJ/m 2
LDPE 150–250 20 300–1000 > 40
HDPE 600–1400 30 100–1000 5–20
PP 1100–1600 30–70 150–700 3–15
PVC 2900–3400 50–80 20–40 2–5
PS 3000–3600 45–60 3–4 2
TPS 1600–2500 20–50 20–50 5–10
SAN 3600 70–80 5 3
ABS 1600–3000 20–50 15–30 8–30
ASA 2300–2600 45–60 15–20 7–14
PMMA 3300 50–80 3–7 2–3
PA 6 1000–2000 35–50 150–250 > 20
PA 6.6 1700–2000 55–60 100–200 15–20
PA 11 1100–1200 40–45 200–250 30–40
PA 12 1200–1350 40–45 100–350 15–30
POM 3000–6000 65–70 15–60 5–8
PC 2000–2200 60–65 80–150 20–35
PETP 2800–3100 55–75 50–150 3–6
PBTP 2600–2800 50–55 100–200 3–6
PPE/PS 2200–2500 50–65 60 > 15
PSU 2450 75–90 50–100 3–10
PPS 3400 75 3
PI 3200 75–90 4–8 4–8
PTFE 450–750 20–40 250–500 14–16
FEP 360 18–22 250–330 > 20
PVDF 800–1800 40–50 50–200 > 20
ETFE 850–1400 30–55 200–400 > 20
CA 1500–3000 50–60 30–40 5–30
CAB 500–2000 35–50 20–70 5–30
PB 450–600 22–25 200–350 > 40
PMP 1500 25–30 15 3–10
PEEK 3700 90 50 55
PES 2440 84 40–80
PK 1500 55 350 20
Data on materials 187

Abbr. Tg Tm Vicat B ISO/A λ α c

°C °C °C °C W/m·K 10 /K kJ/kg·K
LDPE –120 110 55 35 0.35 23 2.4
HDPE –120 130 70 45 0.45 13 2.1
PP –15 170 90 60 0.24 18 1.7
PVC 87 – 85 70 0.16 8 0.9
PS 95 – 90 85 0.15 7 1.3
TPS 90 – 85 80 0.17 8 1.3
SAN 105 – 100 85 0.18 7 1.2
ABS 105 – 100 95 0.17 9 1.4
ASA 100 – 85 85 0.18 9
PMMA 110 – 100 95 0.19 7 1.45
PA 6 50 223 210 85 0.21 10 1.9
PA 6.6 50 260 230 90 0.20 8.5 1.7
PA 11 45 175 170 70 0.27 12 1.4
PA 12 175 170 70 0.23 9 1.2
POM –50 175 165 115 0.29 12 1.5
PC 150 – 145 135 0.21 6.5 1.3
PETP 75 260 170 80 0.24 7 1.0
PBTP 70 210 180 60 0.21 7 1.3
PPO/PS 140 – 130 125 0.23 6.5 1.25
PSU 190 – 185 175 0.22 5.5 1.0
PPS 85 290 135 0.29 5.5
PI > 400 – > 250 0.52 5.5 1.15
PTFE 126 327 110 55 0.25 12 1.0
FEP 290 0.23 9 1.15
PVDF –40 170 130 92 0.12 8.5 1.4
ETFE 270 75 0.24 7.5 1.95
CA 70 210 70 60 0.20 10 1.5
CAB 55 65 60 0.20 12 1.5
PB –25 130 85 60 0.23 12 1.8
PMP 130 240 165 0.17 12 2.2
PEEK 143 334 160 0.25 4.7
PES 230 226 210
PK 15 220 205 100 0.27 11 1.8

Tg = glass-rubber transition temperature

Tm = meting point
Vicat B = softening T at bij 10 N
ISO/A = heat deflection T at 1.85 MPa
λ = heat transmission coefficient
α = linear expansion coefficient
c = specific heat
188 From polymers to plastics

Abbr. water resistance against

adsorption acids alkalis oils and
(ASTM) weak strong oxyd. weak strong fats weather *
LDPE < 0.01 ++ ! –– ++ ++ + !
HDPE < 0.01 ++ ! –– ++ ++ + !
PP 0.02 ++ ! –– ++ ++ + !
PVC 0.04–0.4 ++ ++ ! ++ ++ + ++
PS 0.03–0.1 ++ + – + + + –
TPS 0.05–0.6 ++ ! – ++ ++ ! !
SAN 0.2–0.3 ++ + – ++ ++ + !
ABS 0.2–0.4 ++ ! –– ++ ++ + !
ASA + ! –– ++ ++ + ++
PMMA 0.1–0.4 + – ! ++ ++ ++ ++
PA 6 1.3–1.9 –– –– –– + + + +
PA 6.6 1.5 – –– –– + + + !
PA 11 0.3 ! –– –– ++ ++ ++ !
PA 12 0.25 –– –– –– + +
POM 0.25 ! –– –– ++ ++ + !
PC 0.16 + ! – –– –– + +
PETP 0.39 ++ + ! + – ++ +
PBTP 0.08 ! – ! ++ ++ ++ +
PPO/PS 0.06 ++ ++ ++ ++ ! +
PSU 0.02 ++ ++ ++ ++ + ++
PPS 0.02 ++ ! ++ + ++
PI 0.32 ++ ++ ++ – + !
PTFE 0 ++ ++ ++ ++ ++ ++ ++
FEP < 0.1 ++ ++ ++ ++ ++ ++ ++
PVDF 0 ++ + –– ++ ++ + +
ETFE 0.03 ++ ++ ! ++ ++ +
CA 6 ! – –– ! –– ++ !
CAB 1–3 + – –– + ! ++ +
PB < 0.02 ++ ! –– ++ ++ + !
PMP 0.01 ++ ++ ++ ++ + !
PK 0.5 + – – + – ++ !

* Weather resistance can, in most cases, be improved by proper additives.

Data on materials 189

10.2. Thermosets
d E tensile notched impact ISO/A
kg/dm 3
GPa strength strength °C
MPa kJ/m 2
PF phenolformaldehyde
+ anorg. filler 1.8–2.1 6–15 15–25 2–10 120–170
+ org. filler 1.4–1.5 4–8 25–30 2–10 120–150
UF ureumformaldehyde
+ org. filler 1.5 6–10 30 6–8 110–140
MF melamineformaldehyde
+ anorg. filler 1.8–2 8–13 25 2 150–200
+ org. filler 1.5 6–10 30 2 140–180
UP polyester (normal) 1.2 3.5 30 70
id (high T) 1.2 3.5 55 120
+ short glass fibres 2.0 12–15 30–50 10–20 100–200
EP epoxy (solid) 1.2 4 60 2 90
+ filler 1.8 12 40 2–15 100–200
+ short glass fibres 1.9 15 30–60 10–20 150–220
(fluid resin) 1.2 4 55 2 110–160
+ filler 1.8 12 45 2–10 130–200
(cold-curing) 1.2 4 50 2 40–60
+ filler 1.6 12 35 2 45–90

resin % glass type d E tensile

kg/dm3 GPa strength
polyester 25 cloth 1.35 5 70
polyester 45 cloth 1.45 9 140
polyester 50 cloth 1.60 10 200
polyester 65 cloth 1.80 19 300
polyester 65 unidirect. 1.80 28 500
epoxy 50 cloth 1.60 10 220
epoxy 65 cloth 1.80 18 350
epoxy 65 unidirect. 1.80 30 700
epoxy 78 unidirect. 2.0 60 1700
190 From polymers to plastics

10.3. Elastomers
Abbr. Name density price min. max. tensile strength
3 class temp. temp.
kg/dm unfilled filled
see p. °C °C MPa MPa
NR natural rubber 0.92 2 –50 80 18–30 25–35
SBR styrene-butadiene rubber 0.91 2 –45 100 1–2 18–25
BR butadiene rubber 0.91 2 –70 100 8 17
IR isoprene rubber 0.91 2 –50 80 27 29
IIR butyl rubber 0.92 3 –45 150 18–21 18–21
CR chloroprene rubber 1.23 3 –40 115 21–29 21–29
NBR nitrile rubber 1.00 3 –40 130 4–7 21–32
EPR ethylene-propylene rubber 0.86 3 –50 120 2 14–23
SI silicone rubber 1.10 6 –75 270 2 10–14
SBS thermoplastic rubber 0.95 3 –70 60 20–25 20–25
PUR polyurethane rubber 1.07 4 –20 110 20–40
Fl.R. fluor rubbers 1.85 7 –30 280 15–25 15–25

Abbr. elasti- tear abrasion gass resistance against

. city resist- resist- barrier sun oxid. heat hydrocarbons. acids
ance. ance.
NR ++ ++ ++ ! – + + – !
SBR + ! ++ ! – + ++ – !
BR +++ ! +++ ! – ! + – !
IR ++ + ++ ! – ! + – !
IIR –– + ++ ++ ++ ++ ++ – ++
CR ++ + ++ + ++ ++ ++ + ++
NBR + + + + – + ++ + +
EPR + + ++ ! ++ ++ ++ – ++
SI + – – ! ++ ++ ++ – +
SBS ++ + ! – ! ! - +
PUR + ++ ++ + ++ ++ ++ ! –
Fl.R. + ! + ++ + ++ ++ + ++
Data on materials 191




–100 –50 0 50 100 150 200 250 °C



PA 66

PA 6

–100 –50 0 50 100 150 200 250 °C

Figure 10.1. Modulus of elasticity as a function of temperature for a number of polymers.

192 From polymers to plastics




–100 –50 0 50 100 150 200 250 °C

phenol resins + filler
UP type 2
5 with glass

2 high-T cure


5 UP type 2

2 UP type 1

low-T cure

–100 –50 0 50 100 150 200 250 °C
Figure 10.1. (continued) Modulus of elasticity as a function of temperature for a number of

Processing techniques

11.1. Principles of processing

11.1.1. General
The route from raw material to end-product proceeds in one of more steps, one of
which at least takes place in the fluid condition. Most of the raw materials are
available in the solid state: granules or powders of thermoplasts, moulding powders
or moulding masses of thermosets, bales of rubbers.

In some cases the raw material is already in a fluid state: with thermosets as a resin
for casting or impregnating, with thermoplasts, as an exception, sometimes as a
monomer which is able to polymerise in the mould. In these cases the forming
processes are relatively simple, since they can be carried out on low-viscous fluids
which do not require high pressures to be transported.

Solid raw materials have to be transferred into a fluid or plastic state, which is,
nearly always, attained by heating (in exceptional cases by solution). In the fluid
state the material is shaped, after which the resulting shape is fixed by cooling
(thermoplasts), or by a chemical reaction (curing of thermosets, vulcanisation of

However, this simple schedule, solid - fluid - solid, is, in practice, less simple to be
carried out. First of all the low heat conductivity of polymers plays a role: heating a
polymer mass is a slow process when the heat has to be supplied from the outside.
This is the case with heating by the metal case of a processing machine, or by infra-
red radiation. This disadvantage is circumvented by applying internal friction,
generated in the solid bed of granules, but even more in the highly viscous, just
molten, mass during transportation. Also high-frequency dielectric heating can be
applied to materials which, because of a not too low dielectric loss factor (see §
8.2.2), are suitable in this respect; in particular with welding of film this method is
used, but also in preheating thermosetting moulding powders. A third possibility is
ultrasonic heating, making use of local heat dissipation when sound waves are
damped out in the material. This is a suitable method for spotwelding.

In all these cases heating-up of the polymer proceeds quicker and more
homogeneously. Cooling down, however, cannot be accelerated, and, therefore, takes
194 From polymers to plastics

the larger part of the cycle time with thicker thermoplastic end-products.

A second difficulty is the high viscosity of polymeric fluids, which we already

encountered in § 5.3.1, where we also have seen that the viscosity of a molten
polymer is not a constant. As a matter of fact, the viscosity decreases with increasing
temperature (see Figure 11.2), but this effect is overshadowed by the influence of
shear rate (see Figure 11.1). According to its definition, the viscosity should be the
constant ratio between shear stress and shear rate, but with most polymers these two
quantities are not at all proportional to each other. With increasing shear rate the
shear stress increases considerably less than proportionally, so that the ratio, the
viscosity, decreases (see § 5.3.2). As a result of this phenomenon a tenfold increase
of the throughput through a channel requires an increase in pressure by a factor of
only 2 to 3.

η Pa·s
PMMA 200 °C

10 3

10 2

PMMA 250 °C

1 10 10 2 10 3 10 4 10 5
shear rate sec-1

Figure 11.1. Apparent melt viscosity as a function of shear rate.

Notwithstanding this fact, high pressures are required to transport the fluid mass
through the transport channels or into the mould of a processing machine with an
acceptable speed. Pressure levels of 50 to 150 MPa (500 to 1500 bar) are normal, so
that the machines have to be very robust.

The product, formed via the fluid state is, in some cases, the end-product, but may
also be subject to further shaping as an intermediate product. Examples of the first
case are injection-moulded gear-wheels, compression-moulded gramophone-records,
extruded and on-line blown bottles, etc. In the second case an extruded sheet is, for
instance, subjected to a heat-forming process to produce a light-dome; a gear wheel
may be machined from a compression-moulded block, or an extruded tubular film is
Processing techniques 195

welded to form a bag.

η Pa·s



10 3
120 150 180 210 240 270 300 °C

Figure 11.2 Melt viscosity as a function of temperature.

The wide variety of processing techniques enables the choice between various
possibilities: a thin-walled cup can be made by injection moulding as well as by
vacuum-forming of extruded sheet; a gear-wheel can be made by machining or by
injection moulding or by compression moulding. The choice of the process depends,
i.a., on the requirements set to the end-product; with injection-moulding narrower
tolerances can be achieved than with heat-forming of sheets. Moreover the nature of
the polymer often plays a role; it depends, for instance, on the stability of the
polymer at the high temperatures in a process; the length of the rubbery region on the
temperature scale governs the ease of vacuum forming; a too brittle polymer is
difficult to machine, etc.

A very important factor in the choice of a process is the economy. The manufacturing
costs of the various processing techniques are very different and are strongly
dependent on the number of articles to be made. This point is discussed in the next

11.1.2. Production cost

The cost to produce a plastic article can be subdivided into three contributions:
material cost, machine cost and the cost of the mould.

The material costs are composed of the price of the raw material, the dimensions of
the product and the unavoidable loss of material in the shaping process.

The machine costs comprise the interest of the invested capital, depreciation of the
equipment, energy needed, labour cost, maintenance, accommodation cost, etc. The
196 From polymers to plastics

contribution of the machine costs to the price of the end product is governed by the
cycle time, i.e. the time to produce a single article. Cycle times are strongly different:
depending on the size of the product and the type of process they may vary from a
few seconds up to half an hour.

The mould cost should be considered separately because the mould has to be made
specifically for a single type of end-product. It may vary over a factor of thousand.
This cost has to be apportioned over the number of articles to be made, and,
therefore, it largely governs the economy of the manufacturing process.

The cost to produce a number n of a certain article, can thus be expressed by:

K = A + Bn

in which A is the cost of investment in the mould and the preparation of the
production, and B the cost of material, energy, labour, depreciation etc. The cost per
article is then:

Figure 11.3 shows how, for two different processes, this cost depends on the
production volume; for the first process A is three times as high and B three times
lower than for the second process. The choice is, therefore, strongly dependent on the
envisaged volume of production.




1 2
10 2 3 5 10 3 2 3 5 10 4
number of articles

Figure 11.3. Effect of volume of production on manufacturing cost.

Processing techniques 197

11.1.3. Compounding
Before the actual forming process, the raw material has often to be compounded, i.e.
blended with a number of other components, as mentioned in § 1.1.

Blending can, in a first instance, be carried out in dry condition in the polymer
powder or granules; the degree of blending is, however, not yet sufficient. To obtain
a better dispersion, a blending process in the fluid state is necessary. To achieve this,
various types of mixing devices are available, such as roll mills, internal mixers and
mixing extruders.

Roll mills, in most cases, consist of two heated rolls (Figure 11.4), separated by a
narrow gap, and rotating in opposite directions. The thermoplastic or rubbery
material is, together with the added substances, brought onto the rolls, so that it
transported through the gap by friction. By applying slight differences between the
temperature and the rotation speed of the rolls, the mix is forced to adhere to one of
the rolls, where is it is all the time being cut away by hand an again brought between
the rolls, until a homogeneous blend has been formed. Two-roll mills come in
different sizes; the biggest ones have roll diameters up to 80 cm and lengths up to 2.5

feed hopper

mill rail


guide rail

cut strip

discharge tray

Figure 11.4. Two-roll mill.

Internal mixers, such as the Banbury mixer (Figure 11.5), contain two connecting
chambers, in which blades rotate in opposite directions with a narrow clearance to
the walls, resulting in local high rates of shear. The walls can be cooled or heated.
After mixing, the pieces are removed from the chambers, milled into a sheet and cut
into ribbons for storage or further processing.
198 From polymers to plastics

fill opening

mixing house

mixing chamber

mixing rotor

discharge valve


Figure 11.5. Banbury mixer.

The third type of compounding device is the extruder. Next to its use for fabrication
(which will be dealt with in § 11.4.2), it is applied as a mixer. In essence, it is a
screw pump, in which the mass to be mixed is transported in a heated cylinder by a
screw, or, with twin-screw extruders, by two parallel screws. During this transport
melting and intensive mixing take place. At the end of the screw the blend is pressed
through a number of openings and cooled down; in most cases the strands thus
obtained are on-line cut into granules.

Most manufacturers of end-products do not make the compounds themselves, but

obtain them from the material supplier. In the bigger industries compounding is
sometimes carried out in-house, with the advantages of economy and of flexibility to
tune their products to reach the best compromise of processability on the machines
available and the desired properties of the end-product.

11.2. Casting and compression moulding

11.2.1. Casting
Casting is the simplest processing technique: the mould is filled with a liquid, which,
at ambient temperature, has such a low viscosity that it is able to fill the mould under
its own weight. After filling, a chemical reaction should take place to form the
material. With thermosets this is a reaction between the components which, prior to
filling, have been mixed together, and which react into a three-dimensional network.
Sometimes a catalyst is added to initiate or accelerate this reaction. Dependent on the
combination chosen, curing takes place at room temperature or at an elevated

During curing heat is always being developed, which cannot escape easily due to the
Processing techniques 199

low heat conductivity of polymers (see § 8.1.4). In the centre of the body the
temperature may, therefore, rise unacceptably high and cause thermal degradation.
With casting resins thick parts should, therefore, be avoided. By a proper choice of
the type of material and the dosing of catalysts, the heat development can be

Examples used in practice are polyester and epoxy casting resins, both as two-
component systems. The epoxies are more expensive than the polyester resins, but
they show less shrinkage; the shrinkage, expressed as the volume difference before
and after curing, amounts to 1 to 6 % for epoxies, and to 5 to 10 % for polyesters.

Epoxy casting resins are frequently applied for encapsulation of electro-technical

parts such as coils and transformers. The electrical properties of epoxies are, in
general, better than those of polyesters.

Casting resins may be supplemented with fillers, mostly with anorganic materials
such as quartz powder. Sometimes this is done to lower the cost price, but also, for
some applications, to improve the properties; shrinkage, for instance, is reduced and
hardness is increased.

Some thermoplastics can also be used in casting processes; in these cases

polymerisation takes place in the mould which is filled with the monomer. Examples
are polystyrene, polymerised from styrene, and polymethylmethacrylate from its
monomer. These reactions can take place quite easily in the presence of suitable
catalysts. In this way thick sheets can be produced, but also large articles such as PA-
6 ship propellers or gear wheels, as polymerisation products from caprolactam.

11.2.2. Rotational moulding

Casting of hollow articles is possible when gravity forces are replaced by centrifugal
forces, brought about by rotation. A pipe can be produced by partially filling a
hollow cylinder with liquid, rotating it about its axis at high speed, and heating its
outside wall (Figure 11.6). Rotation about two perpendicular axes enables the
formation of an arbitrarily shaped hollow article. (Figure 11.7).
200 From polymers to plastics

cooling room oven

preheating melting cooling loading/discharge

Figure 11.6. Rotational moulding. Figure 11.7. Rotational moulding with

two rotating axes.

The starting material in these processes is not necessarily in the liquid state; also
pastes of PVC with a plasticiser can be used, and even a well-flowing powder of a
thermoplast. The particles will melt when in touch with the heated wall and will form
a compact layer of material. In these cases the centrifugal force does not necessarily
play a role; the softened powder sticks to the wall and other particles gradually
complete the shaping process. Only the outer surface of an article made this way will
be smooth, which is no objection for many applications.

This method is eminently suited for the manufacture of large objects such as
containers of several m3 content, and also non-closed articles such as furniture,
dashboards etc. The normally applied material thickness is in between 1.5 and 15
mm, dependent on the type of article. The maximum possible thickness is about 30
mm, because with thicker parts the time needed to melt the polymer is so long that
part of the material may already decompose before the whole mass is sufficiently
heated. It is, sometimes, tried to circumvent this difficulty by filling the mould with
pre-melted material, obtained from an extruder.

In the wall thickness fluctuations up to 5 % may occur. As a result of the uneven

temperature in the molten polymer during rotation, and also by the not always
exactly reproducible rate of cooling, deviations in the dimensions of the finished
product may amount to 5 %. Requirements are, that the materials can be molten
completely, that the melt is sufficiently low-viscous, and that the molten polymer
does not degrade too rapidly. Besides plasticised PVC, HDPE and LDPE are often
used, as well as copolymers of PE such as EVA (ethylene - vinyl acetate
copolymer).Because the shear stresses in this process are extremely low, a narrow
molar mass distribution is to be recommended, as discussed in § 5.4. Cycle times
vary between 3 and 40 minutes, dependent on the wall thickness. Cycle times can be
reduced considerably by using machines with multiple moulds, since the cycle time
Processing techniques 201

is predominantly controlled by heating and cooling time. With multi-mould

machines, one of the moulds is heated and filled during the cooling cycle of another
mould (see Figure 11.7).

A very promising development is the combination of the in-situ polymerisation

mentioned before, with the principle of rotational moulding. This method is, in
particular, used for the manufacture of large-sized polyamide articles from the
monomer caprolactam. The big advantage compared with centrifugal casting is, that
one starts off with a low-viscous liquid, and ends with a solid material. During the
whole process the temperature does not exceed the melting point of the polyamide
formed. The article, therefore, does not need a cooling cycle before it can be taken
from the mould; moreover the heating time is much shorter. An example is an oil
tank of 1.200 litres with an average wall thickness of 4.5 mm, which requires 33 kg
of caprolactam. The total cycle time for heating, centrifuging and simultaneous
polymerisation is 13 minutes.

The moulds are mostly made of steel or aluminium. Since the forces during rotation
are relatively small, thin walls of the mould suffice, viz. less than 1.5 mm.

From an economic point of view the centrifugal moulding process is characterised by

the following factors: low cost of machine and mould, relatively long cycle times,
suitability for thin- as well as thick-walled articles. The method is very well suited
for articles made in small series.

11.2.3. Compression moulding

Compression moulding is the simplest technique to transform a raw material from
the solid state into an end-product. The material is, as granules or as a powder,
brought between the two heated halves of the mould, which are then being pressed

Thermoplasts, thermosets and rubbers can be processed by compression moulding.

For thermoplasts this technique is an exception; the best known example is the
manufacture of gramophone records from a copolymer of vinyl chloride and vinyl
acetate (chosen to enable a very good flow into all details of the mould).

Compression moulding of thermoplastics is, in general, unattractive because the

mould has to be heated and cooled in turn; the moulding has to be cooled down
below its glass temperature or melting point before it can be taken from the mould.
With thermosets, compression moulding is commonly applied. The starting material
is a moulding powder, comprising a blend of the resin, the curing agent, and organic
or anorganic fillers such as wood flour, cellulose, textile fibres, asbestos, quartz
powder, glass fibres, etc. These moulding powders are prepared by intensive mixing
202 From polymers to plastics

of the components in mixers, or via impregnation of the fillers with a resin solution,
after which the solvent is evaporated. In most cases a certain amount of pre-curing
takes place during this treatment, which can be adjusted by temperature and time.
The pre-curing results in a reduction of cure time in the mould, and thus in a shorter
cycle time and a better economy of the process. For the same reason the heating time
in the mould can be reduced by preheating the material with the aid of a high-
frequency electric field.

The powder is further heated in the mould; it becomes a fluid and fills the mould
completely, after which the final curing reaction takes place. In this course of events
the flow process and the curing reaction are competitive; a temperature increase
reduces the viscosity but increases the reaction rate. When these effects coincide the
process fails: the viscosity is not sufficiently lowered to allow complete filling of the
mould. If the reaction time is much longer than the time needed for flow, then a
perfect moulding is obtained, but the cycle time is too long. This is schematically
shown in Figure 11.8; the level of the minimum viscosity, ηm, governs the technical
aspect of the process (complete filling, required pressure), while the length of the
time scale, T, is a measure of its economy. By a proper choice of resin formulation
and process conditions, an optimum can be found for each article to be made.




Figure 11.8. Viscosity of a resin as a function of time.

The characterisation of the flow- and curing behaviour of moulding powders is

carried out with the aid of a laboratory device in which the powder is, under a
constant pressure, transported from a pre-heated chamber into a narrow channel. The
distance covered as a function of time is recorded, and from the so-called flow-
diagrams the behaviour of the resin system during curing can be read off (Figure
11.9). In such a diagram the three stages of Figure 11.8 are clearly recognisable.

Moulding powders cure, in general, at temperatures between 140 and 180 °C, under
pressures between 200 and 600 bars. A substantial part of the temperature increase is
Processing techniques 203

provided by the reaction heat, which may raise the temperature to above the
temperature of the wall. This effect will be stronger with thicker articles.

flow path



Figure 11.9. Flow diagram of a resin

Dimension tolerances are important in compression moulding. The dimensions of the

finished article are smaller than those of the mould cavity, in particular because not
only thermal shrinkage but also shrinkage during the reaction counts. When using
resins which develop volatiles upon curing, extra shrinkage occurs due to the escape
of these components.

In its simplest form a compression moulding device is sketched in Figure 11.10. A

somewhat more complicated variant is transfer moulding, where the material is
heated in a separate chamber until it is fluid, and is then rammed into the mould
(Figure 11.11). The moulding time is now shorter, while curing proceeds more
homogeneously. Moreover, the mould cavity is less loaded and is less sensitive to
damage in its details. In fact, transfer moulding is an intermediate stage between
simple compression moulding and injection moulding which will be described in the
next section.

Numerous articles are made by compression moulding, such as bulb fittings,

switches, wall-plugs, crockery, and many other technical and household articles. The
process is very economical, since no expensive machines and moulds are required,
while cycle times are short (the finished article can be taken from the mould without
cooling down).

The dimensions of articles made from moulding powders are limited in two ways:
wall thickness can, in general, not be less than a few mm because of the dimensions
of the powder particles. On the other hand, with too thick articles (a few cm) porosity
in the centre may occur as a result of thermal degradation due to the heat of reaction
and the low thermal conductivity.

Compression moulding of rubbers is not essentially different from that of

thermosets. The starting material is a blend of a rubber, vulcanisation ingredients and
204 From polymers to plastics

fillers, which is heated to become plastically deformable, and is forced to flow into a
mould, where it is kept at a high temperature until the vulcanisation reaction is
completed. Also here, there is a competition between flow and curing; premature
vulcanisation interferes with good processing.

Vulcanisation of rubbers can be carried out as a separate process after shaping.

Shaping can take place at relatively low temperatures since the unvulcanised rubber
has an outspoken plastic nature. It is still solid enough to be transferred from the
mould to a separate vulcanisation chamber, in which, at a much higher temperature,
e.g. by steam, the vulcanisation is achieved.

mould half

moulding powder

filling mould half

expeller pin



Figure 11.10. Compression mould.

Processing techniques 205

mould half
heating/cooling channels

mould half



Figure 11.11. Transfer mould.

11.3. Injection moulding

11.3.1. General
Injection moulding is by far the most frequently applied technique for manufacturing
end-products directly from a thermoplastic raw material. In this process articles are
formed by injecting the molten polymer into a cooled mould. Also on rubbers and
thermosets this method is applied, though at a smaller scale; in these cases the mould
is heated rather than cooled in order to promote the vulcanisation or curing reaction.

In this section we shall mainly consider the injection moulding of thermoplasts.

A thermoplastic material is, usually as a granulate, brought into a cylinder, which is

kept at a constant temperature by electric heating elements. With the simplest type of
machine the material is molten by heat transfer from the cylinder wall. Since plastics
are poor heat conductors, such a heating process takes a considerable time, so that
the risk is run of degradation at the wall.
206 From polymers to plastics

When the material is sufficiently heated, it is transported by a moving plunger under

high pressure (sometimes up to 1500 bar) through a narrow channel into the mould
cavity (Figure 11.12a and 11.12b). In the cooled mould the product cools down
under pressure. After complete solidification the mould is opened and the product is
ejected from it (Figure 11.12c).

The plunger has, in the mean time, reached its original position, so that new
granulate can enter into the cylinder (Figure 11.12c). Since the cylinder acts as a
storage chamber, the cold feed is being heated during the cycle. It is clear that the
high injection pressure would push away the back side of the mould, if this would
not be pressed against the front side by a closing force, which has to be very high, for
instance. for small machines, 10 to 50 tons (100 to 500 kN), for very big machines
up to 10,000 tonnes (100,000 kN).

10 11 12 13
1 central ejecting pin
2 ejecting pin
3 back mould
(moving mould half)
4 mould cavity
5 front mould
6 torpedo
7 cylinder wall
8 heating element
1 2 3 45 67 8 9 9 plunger
10 injection gate
11 injection channel
12 granulate or powder
13 feed hopper

Figure 11.12. Plunger injection moulding machine.

Next to plunger machines, injection moulding machines with screw-plastification

(reciprocating screw machines) are used at an increasing scale. Here the material is
heated by heat transfer from the cylinder wall as well as by internal friction; it is
transported and homogenised by an extrusion screw, and, after reaching the screw
tip, it is injected into the mould by the same screw, now acting as a plunger (Figure
11.13a). After the mould is filled, the transport channel is closed; the screw
continues rotating and transporting molten material to the front side, while pushing
Processing techniques 207

itself backwards (Figure 11.13b). This takes place while the polymer in the mould
cools down sufficiently to be demoulded (Figure 11.13c). Hereafter the mould is
again closed and the cycle is repeated.

injection pressure


back pressure

Figure 11.13. Reciprocating screw injection moulding machine.

Because of the much better temperature homogeneity of the injected material, the
required injection pressure is two to three times lower than in a plunger machine,
under comparable conditions; regions where the polymer is hardly molten cause a
considerable increase in viscosity. Plunger machines are, therefore, only used for
mass production of very small articles. A variant is the combination of screw and
plunger, each in their own cylinder, both of which are discharged into a chamber just
in front of the injection point; the screw forms and transports the melt, which is the
injected into the mould by the plunger.

The mould is a very complicated tool. It comprises two halves: the fixed half
attached to the stationary platen and the moving half attached to the moving platen.
With complicated shapes, or with long flow paths, injection has to take place at
several points. For the transport of the melt to these injection points channels are
present, which are mostly heated (“hot runners”) in order to prevent cooling of the
melt before it enters into the cavity. The mould cavity is surrounded by thick metal
walls, through which cooling channels run (with rubbers and thermosets these
channels are used for heating). Moreover, the mould contains an ejection mechanism
to remove the moulding after opening.

The melt should be present at the appropriate temperature and pressure at each
208 From polymers to plastics

injection point to enable the desired flow pattern into the cavity, while unwanted
phenomena such as poor welding seams and frozen-in orientations or cooling
stresses should be avoided.

Computer techniques have been and are still being further developed to simulate this
complex mould-filling process, so that, for a given geometry and given rheological
parameters and solidification behaviour, it can be predicted. This procedure results in
a considerable improvement of the product quality, since it enables optimisation of
the mould design.

The mould is kept closed during injection by a heavy, hydraulically controlled,

closing mechanism. The magnitude of the required forces can be estimated as
follows: With an injection pressure of, e.g. 1000 bar, the pressure in the mould may,
on average, amount to 1/3 of this value, say 300 bar. When the projected area of the
mould is 500 cm2, the maximum closing force is 1500 kN (150 tonnes). The size of
an injection moulding machine is often characterised by its closing force; for smaller
machines this amounts, e.g. to 200 kN, for bigger ones e.g. 3,000 kN and even up to
100,000 kN. In the latter case the force is generated by eight hydraulic cylinders;
single mouldings up to 170 kg can be made; the machine measures 30m by 13m and
requires an electric power of 1000 kW.

11.3.2. Limitations
A first requirement for a good injection moulding process is, of course, that the
mould is filled well. The pressure at which the molten polymer is injected should not
only be sufficient to transport the melt through the injection channels at the desired
speed, but also to continue this transport within the mould through narrow channels
with cooled walls. During this transport the melt is cooled, so that the viscosity
increases. Eventually, it solidifies, starting-off at the walls. If solidification has
proceeded through the whole cross-section of the channel, then further flow is
impossible, and an incomplete moulding results. The dimensions of the end-product
are limited by this phenomenon; the flow-path, i.e. the ratio of the maximum
distance over which the melt is able to travel and the thickness of the channel, has an
upper value. A practical rule of thumb is, that for well-flowing materials this ratio
should not exceed about 300. When the length/diameter ratio of the object to be
formed is higher, or when the polymer flows less well, then the only solution is to
increase the number of injection points.

As a matter of fact, next to the shape of the mould, the injection pressure and the
temperature of the melt play an important role in the filling process. For a given raw
material, at each pressure level a temperature level can be determined at which the
mould is just filled. This relation is, as far as the polymer is concerned, governed by
Processing techniques 209

the way in which the viscosity depends on temperature and shear rate (see § 11.1.1).
In a T, p diagram these minimum values of pressure form the so-called short-shot
line, which indicates the lower limit in the injection moulding diagram (Figure

T degradation

'short shot' 'flash'

not molten

injection pressure p

Figure 11.14. Injection moulding diagram.

The position of this curve is, of course, governed by the flow behaviour of the
polymer; the fact that the melt-index does not provide sufficient information for this
behaviour, has been discussed in §5.4.

Moreover, an upper limit exists to the p, T conditions, namely pressure levels which
are too high to be withstood by the closing mechanism of the mould. The mould
halves will then be pushed apart, allowing a small quantity of the melt to escape,
which results in a filmy extension of the product (“flash”). At higher temperatures,
so with lower viscosities, flash occurs, as a result of better pressure transfer, at lower
injection pressures. The T, p relation is represented by the “flash-curve” (see Figure
11.14). The working area of the process is completed by a lower temperature limit
which, slightly pressure dependent, represents the condition of complete melting, and
an upper limit for thermal degradation.

11.3.3. Defects
Next to the primary requirement that the mould has to be adequately filled, a number
of other factors, concerning the quality of the end-product, play a role. These
concern, in the first place, the dimensions of the moulding, which should be within
narrow specifications. Shrinkage during solidification always tends to reduce the
dimensions to values, lower than those of the mould. The magnitude of this
shrinkage, if no pressure would be applied, is shown in Figure 11.15. It appears that
for amorphous polymers an average reduction in volume of about 10% occurs when
the polymer is cooled down from processing temperature to ambient temperature; for
crystalline polymers this reduction may even amount to 20 or 25% (Figure 11.15; see
also Figure 3.1).
210 From polymers to plastics

1.4 specific volume



1.2 LDPE



T °C

0 50 100 150 200 250

Figure 11.15. Relation between volume and temperature.



0 500 1000 bar

Figure 11.16. Compressibility of molten polymers.

Solidification, however, takes place under pressure; the cooling melt is compressed
an has thus undergone a volume decrease as a result of its high compressibility
(Figure 11.16), which is able to compensate part of the shrinkage. To demonstrate
the competition between thermal shrinkage and compressibility, combined diagrams
are often used, the p,V,T diagrams, giving the relation between volume and
temperature at several pressure levels. Schematic examples of these diagrams are
given in Figures 11.17 and 11.18, for amorphous and crystallisable polymers,
Processing techniques 211

respectively. These diagrams are valid for a certain, realistic, rate of cooling. For the
amorphous polymers the effect of cooling rate on the glass transition temperature is
of importance, and for crystalline polymers the effect of undercooling. According to
these curves, in both cases the pressure plays an important role (not dealt with in
Chapters 4 and 5): higher pressures increase the glass-rubber transition temperature
and increase the rate of crystallisation at a given degree of undercooling

V 1 bar V 1 bar

1500 1500
bar bar


Figure 11.17. p,V,T-diagrams for an Figure 11.18. p,V,T-diagrams for a

amorphous polymer. crystallising polymer.

Superficial inspection of Figures 11.15 and 11.16 shows that the compensation of
shrinkage by compressibility is not complete. Further reduction of shrinkage in an
injection moulding cycle is reached by maintaining the injection pressure during
some time on the mould cavity, the so-called after-pressure. While the outer layers
solidify and shrink, an extra amount of melt is pressed into the cavity, which further
reduces the overall shrinkage. When, finally, the injection opening is closed (while a
small quantity of polymer sometimes flows back), the polymer in the mould is left to
itself during the final solidification process, while the pressure in the cavity
decreases (Figure 11.19). A rapid pressure decrease to zero (a) indicates a
considerable final shrinkage; when the end-pressure is high (b), the after-pressure has
been too high or applied too long; the shrinkage is negative and the moulding is
tightly clamped in the mould. The ultimate shrinkage is, for amorphous polymers,
mostly between 0.3 and 0.7%, and for crystalline polymers between 1 and 3%.

Besides the total shrinkage, shrinkage differences are of importance. The shrinkage
may vary from place to place, dependent on the time scale of solidification. The
simplest example is a solid block, which solidifies rapidly at its outside; the hard skin
formed this way prevents further shrinkage of the inner part. Hence the density
decreases from the outside to the inside, resulting in internal stresses which may
even lead to tearing and void formation.
212 From polymers to plastics

waiting back flow

pressure time after
in filling cooling shrinkage
gate closed open

0 5 10 15 20 25 30 sec

Figure 11.19. Pressure in the mould as a function of time.

Another example is the injection moulding of a pail, where the injection takes place
at the bottom. The most removed part, the brim of the pail, will solidify first, thus
impeding the shrinkage of the remaining part, in particular of the bottom. The bottom
is, therefore, heavily stressed and may spontaneously break in shock loading. A
better construction of the mould can prevent this drawback, namely by making the
bottom not flat but somewhat hollow, so that the stresses easily disappear by a minor
deformation of the bottom.

Another source of uneven shrinkage is a thickness difference in the moulding. A

thicker part cools more slowly, and attains a higher density than a thin part (see
Figure 3.6). These shrinkage differences may give rise to warping of the moulding,
for instance when it contains reinforcing ribs or a transition from a thick to a thinner
part. Also at the opposite side of a thicker part, sink marks may occur as a result of a
higher local shrinkage.

Shrinkage may also be anisotropic, namely when chain orientations are present. As
discussed in Chapter 5, these orientations result from rubber-elastic deformations in
the melt; they are, therefore, found in parts where the melt has been rapidly cooled
under conditions of a high rate of strain. Shrinkage is higher in the orientation
direction than across, so that, for instance, a flat disk, injected from its centre, tends
to deform into a saddle-like shape (warping).

Defects can also occur when two melt streams join together in a mould, forming a
not properly welded interface; sharp internal angles may give rise to stress
concentrations, etc.

All these defects can be avoided by taking the specific behaviour of the polymer into
account when designing the product. A good mould design is also very important,
Processing techniques 213

e.g. as concerns the number and the position of injection points, the cooling, etc.
Besides, the choice of the polymer plays a role, for instance as to the chain length
and -distribution in view of orientations, and the various process variables (injection
pressure and -temperature, after-pressure, mould temperature).

11.3.4. Foam moulding

A special type of injection moulding is the manufacture of foamed articles. The
foaming process takes place in the mould, so that mould filling requires considerably
lower injection pressures and mould closing forces. Very large mouldings can be
made in this way.

An increasingly applied process is reaction injection moulding (RIM). In fact this

resembles the afore-mentioned mould-casting technique, since it starts off with low-
molecular, low-viscous liquids, which in the mould react together into a solid
product. The injection moulding machine is used to inject this blend into the mould
at a high speed, while the pressure is relatively low. Complete filling of the mould is
facilitated by the addition of a small amount of blowing agent, which brings about a
low degree of foaming; the mould closing forces can be relatively small. Large
mouldings, such as micro-porous bumpers for motorcars (e.g. of polyurethane
rubber) can, in this way, be manufactured economically (see further § 11.6.3).

11.4. Calendering and extrusion

11.4.1. Calendering
Calendering is a continuous forming process, used to produce film and sheet of
practically unlimited length. The whole process takes place on a number of rolls:
homogenising, dosing and shaping on the heated rolls, cooling on a number of
cooling rolls. Figure 11.20 presents a schematic view of the calender. The material is
first heated, mixed and kneaded on a mixing mill, and is then fed into the gap
between the first and the second roll, where it is further homogenised, and then
transported to the third and to the fourth role, respectively. The final shaping is
carried out between the last two rolls, through which the material is squeezed. The
formed sheet is then removed from the roll by a take-off roll and transferred to the
cooling device, after which it is wrapped up.

The pressure difference, required to squeeze the plastic matter through the narrow
gaps between the calender rolls, is generated by the calender itself according to the
principle of the friction pump: When a plane moves in its own direction through a
fluid by which it is wetted, then the fluid, adjacent to the plane, will have the same
velocity as the plane. By transfer of impulse, the whole fluid will start moving, so
214 From polymers to plastics

that transport can take place.

shaping cooling winding

Figure 11.20. Principle of a calender.

When the cross-section of the flow channel decreases, the flow will proceed against a
built-up back-pressure. This back pressure, which reaches its maximum at some
distance before the narrowest part of the gap, forces the material through the gap,
while it obtains a higher speed than the circumferential roll speed. A part of the
material supplied by the upper roll, flows back from the location of the pressure
peak, and arrives at the bank in front of the gap, in which it circulates (see Figure

pressure profile

p max
1 p max

polymer x
melt calendered

Figure 11.21. Pressure profile in the gap of calender rolls.

The generated pressures are very high as a result of the high viscosity of the polymer.
When a 50 µm thick PVC film is made, the line pressure may be as high as 8 kN/cm,
so that a 250 cm roll is subjected to a total force of 2000 kN (200 tonnes). This force
results in a bending deformation of the roll, which increases the distance between the
rolls in the middle. Though the rolls are heavily constructed (diameter about 1/3 of
the length), the deviations in film dimensions may pass the tolerance limits. An
allowable thickness deviation of 5% means a maximum bending deformation of 3
Processing techniques 215

µm! A practical solution to avoid this difficulty is, to position the axes of the rolls
with a small angle with respect to each other. This results in an increase of gap width
from the centre to the outer parts, so that the effect of roll bending is compensated.

Another difficulty with calendering is, that the, still plastic, sheet has to be taken-off
from the final shaping roll. This requires a certain force to overcome the adhesion
between the sheet and the roll. The plastic material should be able to cope with this
tensile stress, and, therefore, should be in a rubbery state. As a consequence, only
polymers showing a pronounced rubbery region, can be processed by a calender.
Possibilities are: PVC (as such or plasticized), ABS, and, as a matter of fact, rubbers.
The elastic stresses generated by take-off are partly frozen-in upon cooling. In all
products made by calendering, therefore, frozen-in orientations are present, resulting
in anisotropy of properties (the “calender-effect”). Moreover these frozen-in stresses
may lead to shrinkage in the length-direction, which, in particular, is a nuisance in
welding operations. A measure to prevent the calender effect is the application of a
strong, flat stream of heated air, which scrapes the sheet from the last roll.

11.4.2. Extrusion
With extrusion, also a continuous process, the material is heated, molten,
pressurized, and forced through a die, all along one and the same screw. The pressure
build-up again takes place according to the principle of the friction pump. In Figure
11.22 the extruder is schematically represented. Its most essential part is the screw,
which fits in a cylinder, provided with heating elements and cooling channels, so that
a desired temperature profile can be established and maintained. At one end of the
screw the filling opening is located for the raw material supply; at the other end the
molten mass leaves the machine via the shaping die to be cooled thereafter.

die molten polymer feed hopper

heating/cooling jackets

metering compression transport section

section section

Figure 11.22. Principle of an extruder.

The screw performs the following tasks:

– transport of the granulate or powder, supplied from the hopper, to the heating
216 From polymers to plastics

– densification of the material by reduction of channel depth or pitch;

– perform work of friction on the, already plastic, mass, for homogenising and
heating (an important part of the required heat is generated by friction);
– pumping the molten polymer to the die and generating the pressure required for
transport through the die.
This multitude of functions makes it impossible to process a variety of polymers with
one and the same screw. Each polymer has its own demands, concerning, i.a.,
granule size, required compression ratio, melting behaviour, risks of degradation, etc.

The final part of the screw acts as a friction pump. This can be visualised most easily
by first imagining the screw at rest in a rotating cylinder, and then unrolling the
screw channel into a straight channel with a rectangular cross-section. The cylinder
wall then forms a lid on the channel, moving at a certain angle. In a cross-section
through the length direction of the channel, a velocity distribution would occur as
sketched in Figure 11.23a, if the end of the channel would be open, namely a linear
distribution between the limits 0 (channel bottom) and v (the component of the
cylinder speed in the length direction). With a flow restriction at the end of the
channel (the shaping die), a back-pressure p is built up. The velocity profile as a
result of this back pressure only, would look like Figure 11.23b: a parabolic
distribution in the opposite direction. The combination of these two distributions is
sketched in Figure 11.23c; a net flow results, which governs the throughput of the
extruder. This throughput, Q, can be estimated as follows: The drag flow to the right,
Qd , is proportional to the velocity, v, and thus to the rate of rotation of the screw, n,
and is also dependent on the geometry of the screw channel, but not on the viscosity
of the melt, so: Qd = a · n. The pressure flow in the opposite direction, Qp, is
proportional to the pressure p, inversely proportional to the viscosity η, and, further,
again dependent on the screw geometry, so :
Qp = b·

De net flow, Q = Qs – Qp , iis the flow transported through the die under the influence
of the pressure, p, and is thus proportional to p/η, and, further, dependent on the flow
resistance of the die. So we see:
p p
Q = an – b =c
η η
Processing techniques 217

a b barrel c

drag flow pressure flow superposition

Figure 11.23. Velocity distribution in the screw channel of an extruder.

rate of
throughput low die resitance

high die resitance

high screw speed

low screw speed

back pressure p

Figure 11.24. Extrusion diagram.

A plot of the throughput Q as a function of the pressure p is given in Figure 11.24.

For a given screw geometry the first equation, the screw characteristic, is given for
two values of the rate of rotation, n, while the second equation, the die characteristic,
is shown for two levels of the flow resistance in the die, 1/c. Each combination gives
rise to an intersection of the two lines, the working point of the extruder, from which
the values of Q and p can be read. Combination of the two equations results in:
Q = nb + c

in other words: the throughput rate is, for a given screw/die combination,
proportional to the rotation rate of the screw, and independent of pressure and
viscosity. This treatment is only valid for a Newtonian fluid (constant viscosity). In
reality, polymers deviate from this simplified picture, though the general conclusions
are not affected. (see also Qu. 11.19).

Besides the single-screw extruder, also twin-screw extruders are being used. Two
218 From polymers to plastics

parallel screws rotate within the same cylinder housing, either in the same or in
opposite direction; in the latter case the screws intersect partially. The mechanism of
transportation is much more complex than with a single screw. Twin-screw extruders
are used for powders which are difficult to transport, but mainly as mixing devices in
the compounding industry. Their mixing and homogenising capacity is much greater
than of a single-screw extruder.

After leaving the die, the extruded strand shows an instantaneous increase in
thickness. This is a consequence of the elastic elongation, in particular induced by
the elongational flow in the die, which spontaneously recovers when the material
leaves the die. This so-called “die swell” or “extrudate swell” has to be taken into
account when designing the die; in particular, when the cross-section of the extrudate
is not circular, not only a diameter increase but also a distortion in shape occurs. The
distortion can be prevented by adjusting the shape of the die opening.

A second consequence of the elastic behaviour of the melt is the occurrence of melt
fracture. At high rates of extrusion the elastic deformation of the melt may become
so high that it locally breaks, resulting in an irregularly shaped extrudate or in
surface roughness.

The range of products which can be made by extrusion is very large: rods, threads,
pipes, hoses, sheets, films, profiles, wire coatings, etc. Threads and rods can be
formed quite easily by applying a round die opening. Pipes and hoses require an
annular opening; after leaving the die, the product is fed into the calibrator, a
somewhat wider cylinder, in which the pipe, either by internal pressure or by vacuum
suction through narrow openings, is pressed against the highly polished inside

Sheet extrusion requires a broad die, in which an even distribution of the flowing
polymer along its width, and an equal flow resistance at each point, has to be taken
care of. Sheets are being extruded up to a width of 3 or 4 m, in thickness up to 3 cm.
If desired, sheets can be produced with a wave-shape, by passing the, still plastic,
extrudate between two profiled rolls. Film extrusion can be done in a similar way, as
a matter of fact with a narrower die gap. After extrusion, sheets and films are cooled
in a water bath or on polished cooling rolls.

An important application of extrusion is the coating of metal wires or profiles with a

plastic layer. This is done on a large scale in the manufacturing of electrically
insulated wires and cables. For this purpose a cross-head is needed (Figure 11.25).
The same principle is used for plastics profiles containing a metal core. These may
have a very complex shape, such as for windowframes.
Processing techniques 219

Figure 11.25. Cross-head.

Rubbers are also being extruded, in a not essentially different way from plastics.
Cooling of the extrusion cylinder is necessary to prevent premature vulcanisation as
a result of the heat developed by internal friction. The extruder is fed by ribbons,
obtained from milled sheets. End products are: hoses, profiles, and cable mantles.
On-line vulcanisation can be achieved by passing the extrudate through a steam
channel, while the rate of extrusion is adjusted to the rate of curing. For this purpose
high-rate vulcanisation recipes have been developed. Steam temperatures of about
200 °C are being applied (15 bars pressure). Treads for motorcar tyres are also
extruded; they are wrapped round the pre-formed carcass and then formed and
vulcanised in a press.

11.4.3. Film blowing

An alternative way of producing film is the blowing process. An annular die is used,
from which the material emerges as a thin-walled tube, which is immediately blown-
up by internal pressure to a much larger diameter (Figure 11.26).

The tubular film thus formed is laid flat and reeled, and can, by cutting an welding,
easily be formed into plastic bags.
Multi-layer films can be produced, composed of different polymers. For each
polymer a separate extruder is applied; the extruders discharge into an accumulator,
in which the separate streams are guided to their own position in the film. A
laminated film can thus be made, in which one of the layers provides the strength,
another one acts as a gas barrier, other layers promote adhesion and weldability, etc.
This laminating process is applied with flat film extrusion as well as with film
Extruded films are sometimes bi-axially stretched, in particular PE and PP films,
which gain in strength and transparency by biaxial orientation. The stretching takes
place at elevated temperatures, but below the melting point. The edges of the film are
gripped by a series of clips, which diverge in the transverse direction and
simultaneously accelerate in the machine direction.
220 From polymers to plastics

take-off rolls

winding device
guide rolls



extrusion die


air supply

Figure 11.26. Film blowing.

11.4.3. Bottle blowing

Blowing of hollow articles such as bottles and containers can be carried out on-line
with the extrusion process. A thermoplastic tube, emerging from the extruder die, is,
in the plastic condition, blown-up against the inside of a mould comprising two
halves, the walls of which are cooled. The neck of the bottle is calibrated with the aid
of a mandrel; the bottom is sealed by the mould parts. The principle is indicated in
Figure 11.27. Since, especially for thick-walled articles, the cooling time largely
governs the cycle time, long waiting times are avoided by using a series of moving
moulds. When the bottle has cooled down sufficiently, the mould is opened, and
moves back to the extruder to be filled again. Since the pressures required for
blowing are relatively low, the mould does not need to be heavily constructed, and is,
therefore, much simpler than the mould in an injection moulding machine.
Processing techniques 221




mould half

air supply

Figure 11.27. Bottle blowing.

The polymers mostly used fo bottle blowing are PE, PVC, PA and PC. The size of
products has gradually increased: bottles are being made up to 60 litres, containers
up to 10 m3 . The blowing process, though simple, has a few disadvantages: the wall
thickness is uneven, pinching-off at the extremes causes a considerable loss of
material, and the weld at the bottom is defacing. These drawbacks are partly taken
away by a variant of the process, the injection moulding blowing process. Here a pre-
form is made by injection moulding, comparable to the extruded parison, but now
much more adjusted to the ultimate shape of the bottle as to shape and wall thickness
distribution. This pre-form is, together with the mandrel round which it has been
shaped, transported to a second part of the mould, in which it is blown-up under
pressure. This process is applied for demanding articles, such as POM aerosol

A newer development is the manufacture of bi-axially oriented bottles, in particular

for PETP. The pre-form is injection- moulded and rapidly cooled in the mould; it
remains amorphous (PETP crystallises very slowly); by heating above its Tg (65°) it
passes into the rubbery state, and can then be blown-up and, simultaneously,
longitudinally stretched. The biaxial orientation thus obtained, accelerates the
crystallisation, and, at the same time, results in a very fine crystalline texture, so that
a thin-walled, strong, transparent and heat-resistant bottle is obtained.
222 From polymers to plastics

11.5. Secondary shaping

11.5.1. Thermoforming
The processing techniques dealt with in this section are all within the category
“secondary shaping”. The starting material is obtained from a primary shaping
process as an intermediate, for instance as a sheet, as a film, a block or a pipe.

Sheet forming is carried out on a sheet, in most cases extruded, sometimes

calendered. The sheet is first heated to above its softening temperature, then formed
and thereafter cooled. The technique most frequently applied is vacuum forming; the
force required for deformation is brought about by a vacuum below the heated sheet,
which sucks the sheet onto the mould. In its simplest form, this process is sketched in
Figure 11.28. Heating is mostly achieved with infra-red radiation sources, cooling
with compressed air or water sprays.


plastic sheet clamping



Figure 11.28. Vacuum forming.

An important factor with vacuum forming is, that the heated polymer is not too soon
passing into the fluid state; the sheet would then, during transportation to the mould,
be deformed too much by sagging-out. Hence, the polymer should show a
pronounced rubbery region on the temperature scale. For this reason, amorphous
polymers such as PVC, PS, PMMA, PC, ABS, etc. are eminently suited for vacuum
forming. With semi-crystalline polymers the rubbery region is largely masked by the
crystallinity above the glass-rubber transition (see Figure 11.29). With PE and PP,
vacuum forming is, therefore, a critical operation, in which the processing conditions
should be very carefully controlled. Only high-molecular grades show sufficient
rubbery behaviour above their melting point to enable a sheet-forming process. PA,
PETP and PBTP, however, cannot be shaped this way.
Processing techniques 223

log E
amorphous thermoplast

use extrusion

vacuum forming


log E
Tg Tm
crystalline thermoplast

'cold' forming

blowing injection

vacuum forming temp


Figure 11.29. Temperature regions in which various processing techniques can be applied.

With thicker articles or with complex shapes, vacuum does not always provide a
sufficiently high force to bring about the required deformation. Therefore, air
pressure is sometimes applied, or plugs, either simultaneously or consecutively;
various combinations are being used. Typical areas of application are: cups and trays
for food packaging, articles with a larger area such as lighting ornaments, casings for
apparatuses and machines, kitchen dressers, bath tubs, rowing-boats, etc. For this
type of articles injection moulding is a less suitable technique, due to the
224 From polymers to plastics

unfavourably long flow-paths and the required high mould closing forces. The
moulds for sheet forming are considerably simpler and cheaper than for injection
moulding because of the much lower forces required; for comparable dimensions
their cost is, on average, five times lower. Cast aluminium is excellently suited for
sheet forming moulds; experimental moulds can be made from timber or gypsum; for
small series moulds of epoxy- or phenolic resin are suitable.

Sheet forming processes show two important drawbacks: In the first place, the
pattern of deformation results in a strongly uneven wall thickness of the finished
article. In a cup, the lower brim may, for instance, be two times thinner than the
middle or the upper rim. The second drawback is associated with the fact that
deformation takes place in the rubbery condition. In this condition straining goes
accompanied by chain orientations, which are frozen-in upon rapid cooling. When
heated to above its glass-rubber transition temperature, the article recovers to its
original shape as a flat sheet, but also at lower temperatures it shows a strong
tendency to shrinkage.

11.5.2. Cold sheet forming

With cold-sheet forming also finished products are formed from sheet, but now in the
solid state, at low temperatures, so that considerably higher plug forces are required.
In most cases the rim of the sheet is not clamped, so that during forming its
circumference decreases: the sheet material is pulled in the inside direction. As a
consequence, the differences in thickness of the finished article are much smaller
than with the heat-forming process, namely not more than about 15%. A schematic
picture of the deep-drawing process is given in Figure 11.30.

11.5.3. Forging
Thick-walled articles can be manufactured by forging a sheet, as shown in Figure
11.31. In this way gear wheels, pulleys, etc. can be made.

In comparison with injection moulding, forging has the disadvantage that the article
is not made in a single step from the raw material, but via an extruded bar or sheet.
Forging, however, offers the significant advantage of a shorter cycle time for thick
articles, e.g. above 10 mm. With injection moulding the cycle time strongly increases
with the material thickness, since the cooling time is proportional to the square of the
thickness. This is demonstrated in Figure 11.32 for a disk-shaped PP article.
Moreover, notwithstanding the higher forces required, a forging mould costs only
about 30% of an equivalent mould in an in injection moulding machine.
Processing techniques 225

step 1

step 2

Figure 11.30. Deep-drawing.

Figure 11.31. Forging.

11.5.4. Bending
Bending of a sheet or a pipe is a simple operation; in most cases the object is heated
till it reaches the rubbery state. Sometimes cold bending is possible, e.g. with a thin
PVC sheet to cover a wooden gutter, though at the cost of the mechanical properties.
Thin-walled channels with a large cross-section are easiest made by bending of
sheet; the seams are then welded.

The heating operation is hampered by the low heat conductivity of the material and
by its sensitivity to heat degradation during long exposure times. Sheets up to 3 mm
can be heated from one side; with thicker ones both sides should be heated. The use
of a liquid bath, e.g. glycol, is, when possible, the best method.
226 From polymers to plastics

cycle time
100 injection moulding

50 forging

0 10 20 30 40 mm

Figure 11.32. Cycle time with cold forging compared with injection moulding.

11.5.5. Machining
Plastics can be shaped by machining, though, in comparison to metals, some special
precautions should be taken:

– The low stiffness necessitates rigid supporting and clamping devices; moreover
the forces exerted should not be too high.
– Brittle plastics may present difficulties with stamping and cutting operations.
– In view of the low softening temperature, combined with low heat conductivity,
very effective cooling is required as well as a low cutting speed.
– Absorption of oil or water may affect the dimension tolerances; moreover some
oils and fats may cause environmental stress-cracking.
– Some polymers develop irritating or poisonous vapours upon heating, such as
PVC and, in particular, fluoropolymers.
– Recoil of the material may affect the dimensions; with drilling, for instance, an
over-measure of a few 0.1 mm is desirable.
Machining operations are expensive due their high cost of labour. This does not so
much count if only a small number of products have to be made. For mass-
production injection moulding is more economical, also with respect to the much
lower material losses.

11.5.6. Welding
Welding of intermediate products such as pipes, sheets or films is mostly applied to
articles in the technical sector (pipe installations, tanks etc.) and in the packaging
sector (packaging film).

Since the viscosity of the molten material is always high, the weld does not flow
Processing techniques 227

spontaneously together such as with metals; a good weld is only formed under
pressure. Another important difference with metals is the lower specific heat of the
polymer per unit volume; together with the lower melting temperatures this results in
much lower required amounts of heat, namely about 10% of those for metals.
Moreover, the heat is not easily removed in view of the low conductivity, so that the
material in adjacent parts remains cold. Because of the high thermal expansion,
shrinkage stresses are readily generated during cooling. These stresses may give rise
to the formation of crazes around weld lines in materials which are are sensitive to
crazing. Crazing may also result from stresses originating from differences in
crystallinity or texture, brought about by the local heat treatment.

An often used welding technique is thread-welding; a heated weld thread is pressed

into a V-shaped groove between the sheets to be welded together. Also welds are
being made without the addition of extra material; with mirror-welding a heated
metal plate is used, against which the two parts are pressed at opposite sides. As soon
as the material is in a sufficiently plastic state, the mirror is removed and the parts
are pressed together. With socket-welding the inside of the socket and the outside of
the pipe are heated and welded together. Thick-walled pipes are sometimes welded
by heat of friction; both ends are brought into touch and rotated at a high speed with
respect to each other; when the surfaces are sufficiently plastified, rotation is stopped
and welding is finished by pressure.

Film welding requires special precautions in view of the small thickness. When the
two overlapping ends are heated by an external heat source, the heat is supplied at
the wrong side of the film so that the films are heavily distorted. With very short
electric pulses good welds can be made, though only with thin films (up to 0.2 mm),
and with special precautions to avoid sticking of the film to the clamping device.

It is much more attractive to heat the films from the inside, namely by the application
of a high-frequency electric field. In this case the clamping device remains cold, and
the highest temperature is generated at the surfaces to be welded together. Only
polymers with a not too low dielectric loss factor (see § 8.2.2) are suitable for this
process, such as PVC and PMMA. PE, PP and PS, on the contrary, cannot be welded
in this way.

A modern method of welding is ultrasonic welding. High-frequency and high-

intensity sound waves are passed through the material; the damping of these waves
transforms the sound energy locally into the heat required for welding, while, again,
the clamping device remains cold.

A special problem is met when films, possessing uniaxial or biaxial orientation, have
to be welded. Heating of these films leads irrevocably to strong shrinkage. To avoid
228 From polymers to plastics

this, PE and PP films are often coated. in a co-extrusion process, with a thin layer of
a polymer which can be welded by high-frequency electric fields. The heat is then
only developed in the outer layers which are welded together.

11.5.7. Gluing
The adhesion of a glue to a substrate can take place by three different mechanisms:

– Mechanical anchoring of the glue in the surface. The surface should be rough or
porous and the glue should be able to penetrate into all surface details.
– Absorption of the adhesive into the surface. The glue should then contain a
solvent which swells the plastic substrate. The glue can then penetrate into the
surface, and, after evaporation of the solvent, it adheres well to the surface. As a
gluing agent a solution of the same polymer is sometimes used.
– Adsorption of the adhesive to the surface can, also with smooth and non-swelling
materials, provide a good joint; molecular attraction between both kinds of
molecules effects the adhesion.
Some polymers do not swell or solve or show any adhesion with any type of
adhesive, such as PE and PP, which can, therefore, not be joined by gluing.
However, a very thin surface layer can be chemically modified by an electric
discharge, or by strong etching chemicals, so that adhesion is enabled. Similar
surface treatments are also applied to enable printing with paint or ink.

Polymers play a dominant role in the preparation of adhesives. Three groups can be
distinguished: thermoplastic, thermosetting and rubbery adhesives. Thermoplastic
glues are being applied as solutions or as suspensions in water. Mostly they are one-
component systems. The presence of the solvent or water is a disadvantage, since
their evaporation takes a long time, and gives rise to a considerable volume decrease.
A well-known thermoplastic adhesive is polyvinylacetate, which is used as a solution
or as a dispersion. Another type, polyvinylether, is in fact a highly viscous fluid,
which, by its high viscosity, provides a certain adhesion, but remains sticky.

Thermosetting adhesives are, in general, two-component systems, and may be cured

either at ambient or at elevated temperatures. After the components have been mixed,
the glue has a limited time of application. Phenol formaldehyde, polyester resins and
epoxies are being used; the latter show a very strong adhesion to practically all

Adhesives on the basis of a rubber are applied as watery dispersions, as solvents, or

as solvent-free fluids. Sometimes the rubber is vulcanised after the gluing process,
sometimes it remains uncured. Polymers often used are butyl rubber,
polyisobutylene, and polychloroprene. A more recent development is the use of
Processing techniques 229

thermoplastic elastomers such as SBS, which is, as a hot-melt adhesive, applied in

fluid condition, form a strong, elastic adhesive layer after cooling.

11.5.8. Surface coating

In some cases a decorative or protecting layer is brought onto a plastic article.
Lacquer layers can be applied from a solution; when the lacquer has to be cured, then
the curing temperature is limited by the softening point of the polymer. The
formation of crazes can never be totally avoided; in most cases a decrease in impact
strength to about half its initial value can be expected.

Flexible plastics and rubbers can, as a matter of fact, only be treated with rubber-
elastic lacquers, mainly on the basis of polyurethane, which, moreover, should be
resistant to oxidation, oils, fuel and UV light. Besides, polyurethane lacquers are
often used for several other plastics, such as PVC, polyamides, ABS and glass-fibre
reinforced resins.

Some polymers, such as PE and PP, should undergo a surface treatment, just as with
gluing and printing, before a lacquer can be applied. This treatment may be a short
exposure to a hot, oxidizing flame, an electric corona discharge, or a strong oxidizing

Another way of surface improvement is metallising. Very thin metal layers (up to a
few µm), are deposited by evaporating the metal under vacuum. This method can be
applied to all types of plastics. Metallised films find their application in electro-
technics and as reflectors for radiant heat. In an electric oven aluminium is
evaporated, while the vapour is precipitated on the surface to be treated; in most
cases an extra protective transparent layer is added.

Coating of polymer surfaces with thicker metal layers (10–30 µm) is a much more
complicated operation. Several pre-treatment steps are required: first the polymer
surface has to be modified in such a way that, by a chemical process, a layer of
copper or nickel can be deposited. With ABS use is made of the circumstance that it
is a two-phase system, consisting of a hard matrix in which rubber particles are
dispersed. The rubber particles present at the surface, are etched away, leaving a
rough, porous surface, which offers a good adhesion to the chemically deposited
copper or nickel. Thereafter the application by electrolysis of further layers of other
metals (e.g. chromium) is simple. Also for PP, PMMA and polyamides, methods
have been developed for chemical deposition of the first metal layer.

Metal layers have a decorative function, but they also protect the plastic material
against scratching, abrasion, chemical attack and UV-aging. Moreover, they
contribute significantly to the bending stiffness of an article, since at the surface the
230 From polymers to plastics

stiffening metal layer has the greatest effect on this property.

11.6. Processing of composites

11.6.1. General
For a number of composite plastic materials the processing techniques are hardly
different from those for the single ones. This holds for polymer blends, polymers
reinforced by particles, and polymers containing short glass fibres. Two categories,
however, ask for special attention, namely the composites with continuous
reinforcing fibres, and the plastic foams. The former are shaped by impregnation
processes, with the latter the composite material, the foam, is also created during the
forming process.

11.6.2. Impregnating
Reinforced thermosets are made by impregnating the fibres, present as mats, bundles
or cloths, with a liquid resin system; after total impregnation the resin is cured. In
most cases glass fibres are used. They can be applied as rovings, i.e. bundles of
parallel elementary filaments, each 5 to 10 µm thick. These rovings offer the
possibility to be positioned in various directions, so that they can match the desired
distribution of strength. By applying, for example, in a cylindrical tank twice as
much glass in the circumferential direction than longitudinally, the strength
properties are adjusted to the various stresses brought about by the internal pressure.

Another form is the glass cloth, composed of rovings. Weaving can take place in
various patterns, varying from practically unidirectional to “square weaves” with
equal properties in the two main directions.

Another possibility is the use of glass mats, composed of glass felt, which is
obtained by joining together pieces of fibre with a small quantity of polyester resin.
The fibres are oriented at random. Glass mat is relatively cheap, and is very well
suited for the construction of complex objects.

A good adhesion between the fibre and the resin is important. Usually, the glass
fibres have already been subjected to a chemical treatment by their manufacturer in
order to improve their adhesion. The most often used agents for this purpose are
based on methacrylate-chromium compounds and on organic silicium compounds.

The main function of the resin is, to provide an even stress distribution over the
fibres. The mechanical properties of the composite depend, in the first place, on the
amount of glass and on its distribution over the matrix.
Processing techniques 231

Fabrication can be done in various ways. The simplest method is “hand-lay-up”,

where the resin is brought into the fibre mass by hand. The resin system can also be
brought onto a mould, together with short glass fibres, by a spraying pistols. Further
shaping of the impregnated fibre mass is possible by using atmospheric pressure; this
is done by vacuum suction through a porous mould wall. Air pressure can also be
used (the “air-bag” method). Curing can be accelerated by heating the mould.

A more streamlined form of impregnating is the continuous method, where fibre

bundles are passed through an impregnation bath, filled with the resin system. The
impregnated bundles are then arranged into their final position. This may be in a
sheet or in a profile (pultrusion), but the wet roving can also be wound around a
mandrel for the production of a pipe or a vessel. In this case, by proper choice of the
pitch angles and, with vessels, rotation of the mandrel round various axes, the
distribution of strength can be varied within wide limits. Curing of the objects thus
made, is mostly carried out in hot-air ovens.

Impregnation techniques are suited to produce articles of strongly varying

dimensions. With the hand-lay-up method with open moulds and with spraying
techniques, the dimensions are practically unlimited. Also with the vacuum- and the
air-bag methods few restrictions are being met. The minimum wall thickness is about
1.5 mm with open-mould, vacuum and air-bag techniques. With filament winding
the minimum is about 1 mm. Maximum values of the thickness are, with hot-
pressing and with filament winding, about 25 mm, with hand-lay-up 15 mm, with
vacuum- and air-bag methods about 10 mm and with drawn profiles 8 mm.

Just as with other processing methods, shrinkage has to be taken into account. With
polyesters shrinkage by curing may amount to 6 to 10 vol %. Glass reinforcement
reduces this shrinkage by about a factor of two.

Next to glass fibres, other fibres are being used at an increasing rate, such as carbon-
and aramide fibres. These are much more expensive than glass, but they excel in
mechanical properties. For high-precision construction with these newer fibres, such
as for applications in aircraft and space-craft, so-called prepregs are available: strips
of material in which the fibres are, in perfect parallel position, embedded in the resin.
The resin has undergone a partial curing reaction, so that the strips can be handled,
but are still capable to deform and to be joined together in the final curing operation
at a high temperature and pressure.

So far, it would seem that only thermosetting resins are suitable for reinforcement
with continuous fibres. Historically, this is a logical situation, since impregnating is
easiest with a not too highly viscous liquid, such as a resin , which is still a low-
molecular substance during impregnation. Molten thermoplastics, on the contrary,
232 From polymers to plastics

have such a high viscosity that simple impregnation is not possible.

Over the past few years, however, techniques have been developed to enable
continuous reinforcement of thermoplastics. The simplest way is to put a cloth and a
plastic sheet on top of each other in a heated press and to carry out impregnation
under pressure. More difficult is the forming of an end-product from the sheet
produced; with conventional sheet-forming techniques the position of the fibres will
be distorted in an unacceptable way. As in nearly all processing techniques, the
modern finite-element methods with advanced computers are able to present
solutions to this problem; in principle they can predict the position of the fibres in the
sheet-forming operation, so that optimum reinforcement is realised in the end

11.6.3. Foaming
A plastic foam is a heterogeneous blend of a polymer with a gas. The gas cells are
between 1 mm and 0.1 mm. Foams are made from thermoplasts, thermosets and
rubbers. In all these cases the foam structure is generated in the fluid condition; with
thermoplasts it is fixed by solidification, with thermosets and rubbers by the curing
or vulcanisation reaction.

Dispersion of the gas phase in the fluid polymer can be achieved in two ways,
namely by mechanically whipping the gas into the liquid (as with eggs and cream),
or by generating the gas within the liquid (as with beer).

The production of foamed rubber by whipping a latex is known for a long time.
Rubber mattresses are still being made in this way. The latex is whipped into a
foamy substance and then poured into a mould. The water is evaporated, and the
rubber is vulcanised with the aid of vulcanization ingredients (sulphur etc.) added
before to the latex. The density of a rubber latex foam for mattresses is about 70
kg/m3 .

Another whipped foam is ureum formaldehyde foam. A watery solution of ureum,

formaldehyde and additives, is whipped to a very light foam. When this is dried at an
elevated temperature, the resin is cured into a hard foam of a few kg/m3 , to be
applied in packaging and as artificial snow in shop-windows etc.

The second, much more often applied, method of foam manufacture, is the
generation of gas bubbles within the liquid. This requires the presence of blowing
agents, which exist in two types: physical and chemical ones. To the first category
belong substances which are soluble in the polymer at a high pressure or at a lower
temperature, but which leave the polymer upon pressure reduction and/or
temperature increase (for comparison: carbon dioxide gas escapes from beer when
Processing techniques 233

the pressure is released).

One of the best-known examples of a physical blowing agent is pentane in

polystyrene. The raw material is available as PS beads in which pentane is solved.
The beads are heated by steam to 100 °C. The polymer softens and at the same time
the solved pentane evaporates, forming vapour bubbles which expand the bead
drastically. The foamed beads are then brought into a closed mould, in which they
expand further by heating to about 120 °C, they fuse together, and fill the mould
completely. In this way closed-cell foams are obtained with densities of 10 kg/m3 or
higher. This PS foam is mostly used as packaging material and for thermal

Another possibility is to extrude foam into sheets; the pressure within the extruder
prevents foam formation; only after leaving the die the pentane evaporates to form
the foam. In most cases the extrusion process is carried out with normal PS granules;
in the cylinder pentane or another blowing agent is injected, where it mixes into the
polymer and brings about foam formation after leaving the die.

Besides physical blowing agents, also chemical ones are used. These are substances
which undergo a chemical reaction upon heating, while giving off a gassy
component (compare rising of dough by the reaction products in the fermentation
process). Chemical blowing agents are, i.a., used in PVC, PE, PP, etc. The foaming
process can take place in the usual processing machines, such as extruders and
injection-moulding machines.

The development of gases may also result from the same chemical reaction which
produces the polymer network, with thermosets and rubber foams. The most
important examples are the polyurethane foams, of which two variants exist: the hard
ones and the soft ones. Both are formed by a reaction between a liquid poly-alcohol
and an isocyanate, in which CO2 gas is formed which can act as a blowing agent.
The process proceeds as follows: The fluids are mixed together and are brought into
a mould, in which the reaction starts instantaneously. During the reaction carbon
dioxide is developed, which foams the liquid. The heat of reaction accelerates the
process, so that the substance is foamed-up and cross-linked at a high rate, and
abruptly passes from the fluid into the rubbery state. The process is often carried out
continuously by moving a mixing head over a conveyor belt with raised borders. The
fluid mass covers the bottom and is transported by the belt. At the end of the belt
foaming is complete. In this way a foam block is produced of several m2 cross-
section at a rate up to 10 m/min.

The blocks can be cut into mattresses, into insulating foam sheets, or into arbitrary
shapes. The foam density is about 15 kg/m3 , while the foam has an open structure, as
234 From polymers to plastics

a result of the disappearance of cell walls when the cells are arranged in a tight
packing in a not too highly viscous liquid. The foam consists of a skeleton of elastic
bars, which is an ideal structure for mattresses; air can easily escape and re-enter.

Hard polyurethane foams are made in a similar way. Again polyols and isocyanates
are used, but in a somewhat different composition, giving rise to the formation of a
tighter network. Because the viscosity changes with time in a different way during
curing and foaming, hard polyurethane foams contain more closed cells; the
formation of a completely closed-cell structure is also attainable. The recipe is then
changed in such a way that the amount of CO2 developed is considerably lower,
while a physical blowing agent is added, so far, in most cases, a freon. In view of the
environmental problems (the ozone layer) other blowing agents are being tried out.
The heat of reaction causes the blowing agent to evaporate at a moment when the
viscosity has risen to such an extent that only closed cells can be formed. Hard
polyurethane foams are mainly applied for heat insulation. They can be used as
sheets, but also as in-situ formed foams, e.g. by using spraying guns to cover the
outer side of a storage tank with insulating foam. With insulation of a cold tank a
closed cell structure is of major importance to prevent diffusion of water vapour and
ice formation within the foam. These phenomena can further be prevented by
applying a layer, impermeable to water vapour, at the warm side of the foam. Cavity
walls can also be filled in-situ with polyurethane foam, as well as with UF-foam.

Sheets of hard foam can be covered at both sides by a thin skin, e.g. by metal or
glass-reinforced plastic, to form light and stiff sandwich panels, to be used as
building panels.

A rapidly developing product is the integral foam, consisting of a foam core which
gradually passes into a massive skin. From hard and half-hard polyurethane foams
mouldings of several tens of kg are made by injecting a blend of the reacting
components into a die which is rotating to achieve a proper distribution of the
material. A pressure of less than 2 bar is sufficient to compress the material at the
wall of the mould before solidification. The cycle times are mostly between 5 and 15
minutes. The densities vary between 200 and 900 kg/m3 , depending on the
application. Metal inserts can be placed in the mould to become encapsulated by the

Thermoplastic structural foams are, in most cases, made in an injection moulding

machine. The gas is either formed from a chemical blowing agent, or injected under
pressure into the molten polymer. The melt is injected into the mould until this is
filled for e.g. 60 to 80 %, after which the foaming process completes its filling. The
mould has only to withstand the pressure of the expanding gas (10 to 20 bar); it can
be lightly constructed and the closing forces are small.
Processing techniques 235

Thermoplastic structural foam (e.g. PVC, PP, PPE/PS, PC) is applied at a large scale
for furniture, apparatus casings, motorcar parts, etc. The advantages are, in the first
place, the relatively simple fabrication technique, in particular for large-sized
products. Moreover, the extra high stiffness is of importance. This, in itself, sounds
like a paradox, since the introduction of gas bubbles in the material reduces its
modulus of elasticity. In a global way, it can be stated that for a relatively high-
density foam the modulus is proportional to the square of the density and, for a low-
density foam, proportional to the density. When we consider, however, the bending
stiffness of a foam bar, than this increases with the third power of its thickness. A
foam bar has, therefore, a considerably higher stiffness in bending than a massive bar
of the same weight. In most cases bars and plates are loaded in bending; hence
constructions are, with the same weight, much stiffer when foam is used. For
structural foam this is even more the case: the solid skin has an extra stiffening effect
in bending.

Recommended for further study:

R.J. Young and P.A. Lovell, Introduction to Polymers, Chapman and Hall, London,

F.R. Schwarl, Polymer-Mechanik, Springer, Berlin, 1990.

P.C. Hiemenz, Polymer Chemistry, Marcel Dekker, New York, 1984.

N.G. McCrum, C.P. Buckley and C.B. Bucknall, Principles of Polymer Engineering,
Oxford University Press, Oxford, 1988.

H. Domininghaus, Die Kunststoffe und ihre Eigenschaften, VDI, Düsseldorf, 1988.

Z. Tadmor and C. Gogos, Principles of Polymer Processing, Wiley-Interscience,

New York, 1979.

D.H. Morton-Jones, Polymer Processing, Chapman and Hall, London, 1989.

R.P. Sheldon, Composite polymeric Materials, Appl. Sci. Publ., London, 1982.

A simple introduction:
A.K. van der Vegt, Kunststoffen in het kort, Delta Press, 1994.

Broad survey:
A.E. Schouten en A.K. van der Vegt, Plastics, 9th ed., Delta Press, 1991.

Cross-sections through polymer science and technology:

A.K. van der Vegt, Kijk op Kunststoffen, Beaumont, 1999.

abrasion 142 cold sheet forming 224
acrylonitrile-butadiene-styrene ABS 16 complex modulus of elasticity 112, 127
additives 13 compliance 108
amorphous 52 composite plastics 20
anisotropy 98, 99, 136 compounding 197
antioxydants 157 compression moulding 201
antistatic additives 153 condensation 12
aramide fibre 85 conducting polymers 153
Arrhenius equation 55, 96, 114 copolymerization 40
atactic. 39 copolymers 26, 63
averages 26 crazes 135, 227
Avrami equation 77 creep 82, 103, 104, 108, 120
creep isochrones 123
Banbury mixer 197 cross-link density 50
bending 225 cross-links 49, 61
binodals 164 crystal growth 74
birefringence 100, 156 crystal morphology 79
blending 197 crystal size 71
blends 20, 63 crystalline fraction 78
block copolymers 63, 165 crystalline structure 78
blowing agent 213 crystallization of oriented chains 83
bottle blowing 220
branching 40 damping 126
brittle cracks 131 dashpot 103
brittleness temperature 144 degradation 150
Burgers 105 degree of crystallization 82, 119
burning behaviour 150 die-swell 99, 218
butadiene rubber BR 19, 26, 128 dielectric heating 193
butyl rubber IIR 19, 26, 128 dielectric loss factor 154
diffusion 159
calender 214 dilatometer 75
carbochemistry 12 dipole forces 48, 61
carbon black 143, 176 dispersion 63, 161, 165
Carilon 26 dispersion forces 49
casting 198 ductile failure 132
cellulose acetate CA 18 dynamic mechanial behaviour 110
cellulose acetate butyrate CAB 18
chain conformations 41 effects of environment 156
chain flexibility 47, 60, 68 Einstein formula 45, 177
chain interactions 48, 61 electric conductivity 180
chain length and distribution 26, 31 electric resistance 151
chain orientation 70, 83, 99, 212 electric strength 155
chain regularity 38 entanglements 86, 92
chain structure 22 enthalpy of mixing 162
chemical resistance 158 entropy 71, 97
chloroprene rubber CR 19, 26 entropy elasticity 87
cis 100 entropy of mixing 162
cis- and trans-configuration 39 environmental stress cracking 98, 134, 159
cis- or trans-chains 39, 67 epoxy resin EP 19
co-continuous 173 equivalence of time and temperature 114
coil density 44 ethylene tetrafluoroethylene EFTE 17
238 From polymers to plastics

ethylene-propylene rubber EPR, EPDM 20

extrusion 99, 198, 215 Kelvin-Voigt element 104
Kerner 172, 176, 178
fibres 83, 84, 99 light scattering 34
fillers 13 linear viscoelastic behaviour 108, 109
film blowing 219 liquid-crystalline polymers 84, 119
film extrusion 98 long fibres 181
films 99 long-term strength 131
flash 209 loss factor 127
Flory-Huggins 162 loss modulus 111, 127
fluid condition 92 low-density PE 40
foaming 232 lower critical solution temperature 164
foams 21, 149 lyotropic 85
forging 224
fractionation 37 machining 226
free enthalpy 56 Mark-Houwink relation 36, 45
free volume 53, 160 master curve 114, 121
friction 128, 139 maximum temperature of use 150
fringed micel structure 80 Maxwell element 103
mechanical damping 54
gel permeation chromatography 38 melamine-formaldehyde MF 18
glass - rubber transition 144, 164 melt elasticity 31, 96, 97
glass-rubber transition temperature 52, 53, 187 melt fracture 99, 218
glassy state 52, 53 melt index 37, 93, 97
gluing 228 melting point 67
graft copolymer 40 melting region 70, 82
grip 128 metallising 229
metallocene 39, 71
half-synthetic polymers 11 miscibility of polymers 161
hand-lay-up 231 modulus of elasticity 102, 117, 118, 172, 186.
hardness 138 189
heat-build-up 128 mol mass distribution 97
heterogeneity index 29, 30, 33 molar mass 26
heterogeneous nucleation 73 molar mass distribution MMD 31, 37, 97,
high-density PE 40 9995
high-frequency electric field 227 Mooney viscosity 93
homogeneous blend 165
homogeneous nucleation 73 natural polymers 11
Hooke’s law 102 natural rubber NR 67, 100
hydrogen bridges 49 networks 14
hysteresis work 127 Nielsen 173
nitrile rubber NBR 19
impact strength 54, 98, 128, 136 non-linearity 122, 126
impregnating 230 non-Newtonian behaviour 93, 95
in-situ polymerisation 201 nontransparent 83
injection moulding 98, 205 notched impact strength 136, 186, 189
integral foam 234 nucleus formation 72
interaction forces 68 number average M n 27, 31, 33, 98, 99, 133,
interpenetrating network 175 138
intrinsic viscosity 35 nylon 16, 66
ISO Heat Deflection Temperature (HDT) 145,
187, 189 optical anisotropy 156
isoprene rubber IR 19, 26, 100 orientation 78, 83, 100
isotactic 39, 66 osmometry 34
Index 239

oxygen index 150 pultrusion 231

permeability 159
petrochemistry 12 random walk 41, 86
phenol-formaldehyde PF 18 random-copolymer 40
physical ageing 59, 124, 126 rate of crystallization 67, 74
plasticizing 62 reciprocating screw injection moulding
Poisson 113 machine 207
polyacetylene 22 refraction index 83
polyacrylonitrile PAN 25, 61, 67 reinforcement 20
polyamide PA 16, 66, 69 reinforcement by particles 176
polybutadiene BR 19, 22, 26, 39, 67 reinforcing fillers 143
polybutylene PB 18, 25, 62, 67 relaxation spectrum 107
polybutylene terephtalate PBTP 17, 24 relaxation time 97, 104
polycarbonate PC 16, 23, 60, 66, 99, 118 retardation spectrum 108
polychloroprene CR 26 roll mills 197
polyether ether ketone PEEK 18, 24 rotational moulding 98, 199
polyether sulphone PES 18, 24 rubber-fluid transition 91
polyethylene PE 15, 22, 60, 65, 68, 98 rubbery phase 86, 100
polyethylene terephtalate PETP 17, 23, 60, 66, rubbery region 53, 222
polyethyleneoxyde PEO 23 SBS 20, 167
polyethyleneterephtalate 66 screw-plastification 206
polyimide PI 17, 25, 60 secondary transition 54, 127, 137, 144
polyisobutylene PIB 25, 65 segregation 165
polyisoprene IR 19, 26, 39, 67 shear modulus 102
polyketone PK 18, 26 Shore hardness 138, 145
polymer blends 161 short fibres 177
polymer synthesis 12 short-shot 209
polymethylene pentylene PMP 18 shrinkage 209
polymethylmethacrylate PMMA 16, 25 silicone rubbers 20, 24
polyoxymethylene POM 16, 23, 65 size-exclusion chromatography 38
polyphenylene ether PPE 17, 24, 63, 161 softening 145
polyphenylene oxide PPO 17 spectrum 97, 120, 125
polyphenylene sulphide PPS 17, 24, 60, 65 spherulite 81
polyphenyleneether 65 spinodal 163
polyphenylenesulfide 60, 65 stereospecific 39, 66, 67
polypropylene PP 15, 25, 60, 66, 68 steric hindrance 47
polypropylpene 25 storage modulus 110, 127
polystyrene PS 15, 25, 60, 68 stress 103
polysulphone PSU 17, 24, 65 stress corrosion 158
polytetrafluorethene 68 stress relaxation 107
polytetrafluorethylene PTFE 17, 25, 65 stress-strain diagram 117
polyurethane PU 19, 23, 233 styrene butadiene rubber SBR 19, 167
polyurethane rubber PUR 20 styrene-acrylonitrile SAN 16
polyvinylalcohol PVAL 66 superposition principle 108, 123
polyvinylcarbazole PVK 60 surface coating 229
polyvinylchloride PVC 15, 25, 61, 67 syndiotactic 39
polyvinylidene fluoride 17 synthetic elastomers 14, 19
polyvinylidenechloride PVDC 25, 65
polyvinylidenefluoride 17, 25 Takayanagi 172
power-law 95 Taylor equation 169
prepregs 231 tear strength 99
production cost 195 teflon 65, 68
pseudoplastic behaviour 94, 96 temperature-index 150
240 From polymers to plastics

tensile strength 130, 189, 190

tetrafluoroethylene-perfluoropropylene FEP
thermal conduction 148, 187
thermal expansion 147
thermal properties 144
thermoforming 222
thermoplastic elastomers TPE 20, 63, 118, 168
thermoplastics 14, 15, 185
thermosets 14, 18, 189
thermotropic 85
theta-solvent 45
time and temperature 56
time scale 91
time-temperature shift 114
torsion pendulum 112
transparency 155

U.L. index 150

ultrasonic welding 227
unsaturated polyesters UP 18
upper critical solution temperature 164
ureum-formaldehyde UF 18
UV stabilizers 157
vacuum forming 100, 222
Vicat Softening Temperature (VST) 145, 187
visco-elastic behaviour 91, 97, 102
viscosity 92, 103, 194
viscosity average 35
volume retardation 59

wateradsorption 188
Weber number 168
weight average 27, 34, 92
welding 226

yield point 117

z-average 28, 31


1. Introduction
1.1. Vegetable: timber, cotton, cork, sisal, hemp, straw, reed, cane, capok, rattan,
etc. Animal: wool, silk, fur, mohair, horse-hair, intestine, down, ivory, horn, sponge,
beeswax etc.

1.2. With half-synthetic plastics the polymer chain has been formed in a living
tissue, but it has been chemically modified afterwards.
Cellophane is, after chemical modification, obtained from the cellulose in wood, just
as paper (from cellulose and lignin), cellulose fibres (“rayon”), and cellulose plastics.
Leather is made from animal hides in a tanning process.
Natural rubber has to be vulcanised (with sulphur) to obtain a technically usable
Ebonite is also made from natural rubber, but with a much higher amount of sulphur;
it belongs, therefore, to the family of thermosets.
Linen is obtained from flax after a chemical treatment (retting).
The boundaries between half-synthetic and natural polymers are sometimes vague.

1.3. In a cracking process the long, saturated, hydrocarbon chains in the crude oil are
transformed into shorter ones, which, for lack of hydrogen atoms, contain double
bonds, (-C=C-), which provide polymerisability.

1.4. When a diene, C=C-C=C, is polymerised, only one of the two double bonds is
used; an unsaturated chain results, -C-C=C-C-, which offers the possibility of
vulcanisation with sulphur.

1.5.a. 1.4-polybutadiene, -CH2 -CH=CH-CH2 -

b. 1.2- of 3.4-polybutadiene, -CH-CH2-CH-CH2-
| |
|| ||
in which under a. two different structures, cis and trans can be formed.
-C C- /
\ / or C=C
C=C /

1.6. Ethylene-propylene copolymers, and also polyisobutylene, are rubbers with

saturated chains, which cannot be vulcanised with sulphur. They are, therefore,
2 Problems Polymers

copolymerised with a small amount of diene: EP with a diene in a sidegroup (e.g.

dicyclopentadiene), PIB with isoprene (IIR = isobutylene-isoprene rubber = butyl

1.7. a. Polycondensation is possible with saturated monomers, e.g. a diol with a

bifunctional carbon acid, by ester formation on two sides. In this reaction a small
molecule, e.g. water is also formed.
b. After ring opening of saturated monomers these can also polymerise (e.g.
caprolactam to PA-6).

1.8. a. By a reaction between tri-functional and bi-functional components, such as

with the formation of phenol-formaldehyde or a polyurethane from glycerol and di-
b. By cross-linking polymer chains containing double bonds: with rubbers using
small amounts of sulphur; with unsaturated polyesters using (poly)styrene.

1.9. “Thermosets” are cured into networks at elevated temperature from small
molecules or from single chains.
The name could suggest that thermosets become harder at temperature increase; on
the contrary: they soften just as all polymers at their glass-rubber transition, though
their stiffness remains much higher than that of a rubber.
Moreover: curing is not always carried out at a high temperature; also “cold-curing”
resins are used (e.g. in two-component adhesives or in large ship decks).

1.10. In thermoplasts the parts of different chains are bound to each other by weak
secondary bonds, which are broken at higher temperatures. In thermosets and
vulcanised rubbers the chains are permanently linked by primary chemical bonds,
which, also at higher T, do not allow flow.

1.11. Thermosets have a tight network; with rubbers it is much looser; the cross-
links are just sufficient to connect all chains at a few points.

1.12. In thermoplastic rubbers the chains are not connected by primary bonds as is
the case with sulphur-vulcanised rubbers, but by weaker secondary bonds (e.g.
within a PS domain in SBS), which are loosened upon temperature increase).

1.13. Thermoplastic rubbers (TPE’s) show rubbery behaviour at ambient

temperature because the rubber phase plays the dominant part. In thermoplastics the
behaviour of the glassy phase or of the crystalline part is dominant.

1.14. – Thermoplasts: PE, PS, PVC, PP.

– Thermosets: PF, UP.
– Rubbers: SBR.
Answers 3

1.15. LDPE, HDPE and LLDPE. They differ in density: LDPE between 0.91 and
0.92, HDPE between 0.94 and 0.96, and LLDPE between 0.92 and 0.95 g/cm3. Since
the density reflects the crystallinity, they also differ considerably in stiffness,
between as well as within types.

1.16. – thermoplasts: PS, TPS, SAN, ABS en ASA.

– Thermosets: UP (is cross-linked with styrene).
– Rubbers: SBR and SBS.

1.17. PTFE with its very high melting point (327 °C) can only be transformed into
an end product as a powder in a sintering process, since as a melt it would degrade at
the high temperatures involved in other processing operations. Other fluor polymers
such as FEP, PVDF and EFTE are suitable for the usual processing techniques.

1.18. PI, polyimide, can be used short-time up to 500 °C. Related to PI are: poly-
etherimides, polyesterimides en polyamidimides.

1.19. a. a. Blends of polymers

b. Particle- and short fibre- reinforced polymers
c. Composites
d. Foams

1.20. a. Carbon black is added to thermoplasts to decrease the electric resistance, in

particular to prevent electrostatic charge, and also to shield the polymer from
light, thus preventing ultraviolet degradation. Carbon black is added to rubbers to
increase strength and abrasion resistance,
b. Chalk is a reinforcing filler in thermoplasts and thermosets.
c. Glass fibres are added to thermoplasts and thermosets to increase stiffness and
strength, more than particles do.
d. Plasticisers are added to a polymer to reduce the glass rubber transition
temperature drastically (e.g. with PVC), so that the polymer behaves as a rubber
at ambient temperature rather than as hard glassy thermoplast.
e. Mica is a reinforcing filler which, due to its platelet structure, increases the
stiffness more than spherical particles do.
f. Accelerators are added to thermosets and rubbers to speed-up the curing or
vulcanisation process to such an extent that an end product can be formed under
optimal conditions.
g. Lubricants are sometimes used to reduced the friction resistance in a shaping
h. Sulphur is essential in the usual processing technology of rubbers; it brings about
the formation of “cross-links”, which provide the necessary shape stability.
i. Pentane is a blowing agent, applied in a.o. PS, to form foam.
4 Problems Polymers

j. Wood flour is a cheap filler, which brings about some stiffening (e.g. in PVC),
and which reduces the price.
k. Antioxidants can, in the first place, protect a polymer against degradation at the
high temperatures during processing; in the second place they can (mostly
different types) prevent chain rupture during long-term use at elevated
l. SiC is a very hard filler, mostly used to increase the abrasion resistance of a
m. UV-stabilisers protect a polymer against chain break under the influence of the
ultraviolet components of daylight, often in combination with pigments (e.g. with
n. Anti-static agents bring about a very thin layer of a conducting (hydrophilic)
substance at the surface of the product, which allows electric charges, resulting
from friction or contact, to be dissipated quickly.
o. Rubber particles are blended into a brittle polymer to increase its impact strength,
e.g. butadiene rubber in PS, etc.

2. Molecular composition
2.1. – C-atoms only, or also other ones such as N, O, S or Si,
– saturated or unsaturated chains
– simple chains or chains containing rings, even composite rings (examples in MT

2.2. a. PE -CH2 -CH2 -

polyacetylene -CH=CH-CH=CH-
POM -CH2 -O-CH2-O-
b. PP, PVC, PS, PB, PAN.
c. PIB, PVDC, PVDF (see MT 2.1.2).

2.3. PE, -CH2 -CH2 -, monomer mass = 2C + 4H = 2 ¥ 12 + 4 ¥ 1 = 28 g/mole,

PS, -CH2 -CH(C6 H5 )-, monomer mass = 8C + 8H = 8 ¥ 12 +8 ¥ 1 = 104 g/mole;
PA-12, -(CH2 )11-CO-NH-, monomer mass = 12C + O + N + 23H =
12 ¥ 12 + 1 ¥ 16 + 1 ¥ 14 + 23 ¥ 1 = 197 g/mole.
M(PE) = 500 ¥ 28= 14,000 g/mole,
M(PS) = 500 ¥ 104 = 52,000 g/mole
M(PA-12)= 500 ¥ 197 = 98,500 g/mole.
-C-C-C-C-C- chains lie zig-zag shaped, with C-C distances of 0.154 nm and with an
angle between the bonds of 109.5°. The contribution of a C-C link to the contour
length is, therefore, 0.154 ¥ sin 54.2° = 0.126 nm.
– for PE the stretched length is thus 1000 ¥ 0.126 nm = 126 nm (since the monomer
Answers 5

contains two C-atoms),

– for PS also 126 nm,
– for PA-12: 13 links (for simplicity taking the same length for CO-NH-), so 500 ¥
13 ¥ 0.126 nm = 820 nm.
The calculation is only valid for strictly linear PE, without branches. As to PS, it
holds only for the syndio-tactic structure, since otherwise the neighbouring phenyl
rings prevent a stretched conformation. (MT 2.5).

2.4. In this case equal numbers (both 1) are mixed. The number fractions are both
1/2 and the number average is 12 ¥ 80,000 + 12 ¥ 120,000 = 100,000. The weight
fractions are, respectively, 80/200 and 120/200 or 0.4 and 0.6, so that the weight
average is: 0.4 ¥ 80,000 + 0.6 ¥ 120,000 = 104,000.

2.5. ni = Ni /ÂNi = number of the species i divided by the total number = number

fraction, wi = n i ·Mi /Âni ·Mi = weight fraction, zi = w i·Mi /Âw i ·Mi = z-fractionMn =
— —
Âni ·Mi , Mw = Âwi·Mi , Mz = Âzi ·Mi , Âni = Âwi = Âzi = 1 (all normalised).

2.6. Mn = Âni·Mi ; ni = (wi /Mi)/Â(wi /Mi),

Mn = Â((wi/Mi))·Mi )/Â(wi /Mi)) = Âwi /Â(wi /Mi)) = 1/Â(wi /Mi))
— — —
2.7. Mw = Âwi·Mi ; wi = (n·Mi )/Â(ni ·Mi ) = ni ·Mi M
/ n ; Mw = (Âni·M2i )/Mn ;
— —
Mw·Mn = Âni·Mi .

— — —
2.8. Mz = Âzi ·Mi ; zi = (wi ·Mi )/Â(wi ·Mi ) = wi ·MM 2
i / w; Mz = (Âwi·Mii )/Mw;
— —
Mz·Mw = Âwi·Mii .2

2.9. a. In a mono-disperse polymer all chains are equally long (exception). Then Mn
— —
= Mw = Mz etc..
— —
b. The degree of dispersion or heterogeneity index, D, is defined as Mw/Mn ,, and is,
therefore, D = 1 for a mono-disperse polymer.

2.10. The chains break in a statistical way, by which shorter ones of various lengths
are formed. The molar mass distribution is broadened in the direction of the low-
molecular side.

2.11. In the blend every (wi )A is replaced by: wA·(wi )A and every (wi )B by

wB·(wi )B. So Mw of the blend is:

Mw = Â(wA·(wi )A·(Mi )A) + Â(wB·(wi )B·(Mi )B)
— —
= wA·Â(wi )A·(Mi )A + wB·Â(wi )B·(Mi )B = wA·(Mw)A + wB·( Mw)B

2.12. The number of chains in one gramme of polymer is N/Mn (N = Avogadro
number). One gramme of the mixture contains wA grammes of A, wB grammes of B
and wC grammes of C, so that:
— — — —
N/Mn = wA·N/(Mn )A + wB·N/(Mn )B + wC·N/(Mn )C
6 Problems Polymers

— — — —
of 1/Mn = wA/(Mn )A + wB/(Mn )B + wC/(Mn )C.

The Mn values of A, B and C are 10, 20 and 30 kg/mole respectively 10, 20 en 30

kg/mole; wA = wB = 10/(20 + x), wC = x/(20 + x), Mn = 15 kg/mole, so:
1/15 = 10/(20 + x)·10 + 10/(20 + x)·20 + x/(20 + x)·30, which results in: x = 5.
— —
2.13. Mz = ÂwiM2i /Âwi Mi of Âwi M2i /Mw. Further evaluation is analogous to 2.11 and
— — — — — — —
results in: Mz·Mw = w1 ·(Mz)1 ·(Mw)1 + w2·(Mz)2 ·(Mw)2 . So: Mz = 154.5 kg/mole.

2.14. 0.382 = k·(76,000)a; 0.91 = k·(250,000)a; 2.21 = k·(850,000)a; we can plot

log[h ] versus log M to determine the slope a; we can also consider ratios of [h]:
0.91/0.382 = (250/76)a; a = 0.729; 2.21/0.91 = (850/250)a; a = 0.725; 2.21/0.382 =
(850/76)a; a = 0.727; on average a = 0.727. With 0.91 = k·250,0000,727 we find: k =

2.15. w A = 0.37; w B = 0.18; w C = 0.45. Mw (in kg/mole) = 0.37·76 + 0.18·250 +
0.45·850 = 455.62 = 455,620 g/mole.
— —
1/Mn = 0.37/76 + 0.18/250 + 0.45/850; Mn = 163.457 = 163,457 g/mole.

2.16. c t t – to (t – to )/to (t – to )toc = hred

0 95.0 0
0.1 107.7 12.7 0.1337 1.337
0.2 121.4 26.4 0.2779 1.390
0.3 136.3 41.3 0.4347 1.449
0.4 152.8 57.8 0.6084 1.521

When we plot hred against c, it is clear that extrapolation of hred to leads to 1.30 dl/g,
so [h ] = 1.30 dl/g. From Mark-Houwink it follows (see 2.14): 1.30 =
— —
10.8·10–5·M0.727, from which Mv = 410,000 g/mole, indeed slightly below Mw.

1.55 (dl/g)






0 0.1 0.2 0.3 0.4 c (g/dl)
Figure A.1. Figure to the answer of problem 2.16.
Answers 7

2.17. With Mv = (ÂwiMai )1/a:

Mv = (0.37·76,0000,727 + 0.18·250,0000,727 + 0.45·850,0000,727)1/0,727 = 409,340,
close enough to the answer on 2.16 .

2.18. We first calculate the various averages:

w1 = 0.10, w2 = 0.18, w3 = 0.25, w4 = 0.17,
w5 = 0.12, w6 = 0.08, w7 = 0.06, w8 = 0.04.
M1 = 15, M2 = 27, M3 = 39, M4 = 56, M5 = 78, M6 = 104, M7 = 120, M8 = 153.

Mn = 1/Â(wi /Mi)) = 1/(0.10/15 + 0.18/27 + …) = 38.7 kg/mole.

Mw = Â(wiMi ) = 0.10·15 + 0.18·27 + … = 56.6 kg/mole.
— —
Mz = Â(wiM2i )/Mw = (0.10·152 + 0.18·272 + …)/56.6 = 78.8 kg/mole.
— — —
Mz+1 = Â(wiM3i )/MzMw = (0.10·153 + 0.18·273 + …)/(78.8·56.6) = 99.2 kg/mole

a. The melt viscosity (zero-shear) is governed by Mw, so as far as this is concerned
the blend resembles fraction 4 .
b. The blend has a broader molar mass distribution; it behaves, therefore, stronger
non- Newtonian (see MT 5.3.2 ) and flows better at high rates of shear.

c. The impact strength is mainly dependent on Mn , so the blend resembles fraction 3.
d. The die-swell is a consequence of melt-elasticity which is governed by higher
— —
averages such as Mz and Mz+1 ; here the fractions 5 and 6 are of importance.
e. The Young’s modulus in the glassy state is independent of M.

2.19. In kg/mole: M1 = 60, M2 = 600,Mn =1/Â(wi /M1) = 1/((1 – w2 )/60 + w2 /600) =

100, so w2 = 4/9 and w1 = 5/9; so 40 kg are needed.Mw = (5/9)·60 + (4/9)·600 = 300

kg/mole = 300.000 g/mole. Mz = [(5/9)·602 + (4/9)·6002 )]/300 = 540 kg/mole =
540.000 g/mole.

2.20. If we consider the coil as a sphere with a radius r = (·r 2o Ò)1/2 = 73.5 nm then the
volume of such a sphere is V = (4/3)pr3 = 1.663·10–21 m3. The mass m of the chain is
1000/6·1023 = 1.667·10–21 kg. The density, r , is then m/V = 1.00 kg/m3 . When we
compare this with the density of PS, 1050 kg/m3, then the coil is 1050 times ‘diluted’

2.21. With M = 10,000 ·r 2o Ò = n·b2 is 100 times smaller; so the radius of the spherical
coil is 10 times smaller and the volume V is 1000 times smaller. The chain mass is
100 times smaller, so the density, r, is 10 times greater. The ‘dilution’ is then 105

2.22. In a tightest array of spheres the volume fraction is about 0.65; if we consider a
maximum volume fraction of 0.3 to exclude all interaction effects, we find in the
first case (M = 10 6 ) a maximum concentration of 0.3/1050 = 0.29 g/l = 0.028 g/dl
(the densities of solvent and polymer considering as equal for convenience), and in
the second case ten times 10 higher, so 0.28 g/dl.
8 Problems Polymers

2.23. The slope of h sp versus c is greater as M is higher (at lower concentrations

more effect of entanglements).

2.24. For convenience in kg/mole: wA·200 + (1 – wA)·30 = 81 ; wA = 0.3; wB = 0.7,

so 300 kg of A and 700 kg of B.
— — — —
(Mn )A = 40, (Mn )B = 6, (Mz)A = 600, (Mz)B = 90 kg/mole.
— — — —
1/Mn = wA/Mn )A + wB/(Mn )B = 0.3/40 + 0.7/6; Mn = 8.054 = 8054 g/mole.
— — — — — —
Mz = [w A ·(Mz)A·(Mw)A + w B ·( Mz)B·(Mw)B]/ Mw = (0.3·600·200 + 0.7·90·30)/81 =
467.8 = 467,800 g/mole.
— —
A directly produced batch would show: Mn = 81/5 = 16.2 = 16,200 g/mole, and Mz =

3·81 = 243 = 243,000 g/mole. So the blend, though “on-spec” as regards its Mw, has
— —
a two times lower Mn and a two times higher Mz as a user would expect. A blown
bottle is, therefore, less resistant against stress-corrosion and impact failure at low
temperatures. As a result of the stronger pronounced melt-elasticity the die-swell is
higher , so the bottle is too thick and its manufacture requires more material!

2.25. a. a is usually between 0.5 and 1.0.

— — — —
b. Mv is usually between Mn and Mw, but closer to Mw as a approaches to a = 1.
— — — —
c. When a = 1, Mv = Mw ; when a > 1, Mv > Mw (this can occur with very stiff

2.26. When in solution, both substances will cause an increase in boiling point. At
the same concentration this increase will, for a low-molecular substance, be
considerably higher than for a polymer (readily a factor 100 of 1000 times greater),
since the increase in boiling point is proportional to the number of molecules solved.
An even simpler criterion is the viscosity of the solution: with a low-molecular
substance hardly any change is noticed, where a polymer solution exhibits a
significant increase in viscosity.

2.27. The number average degree of polymerisation P n = 1/(1 – p), in which p is the

degree of conversion. The weight average degree of polymerisation P w =
— — — —
(1 + p)/(1 – p) (see MT § 2.2.2). so the heterogeneity index Mw/Mn = P w/ P n = 1 + p,
close to 2.

2.28. a. HO-C2 H4 -OH = 62; HOOC-C6 H4 -COOH = 166; the mass ratio is 62:166 =
b. HO-C2H4 -OH + HOOC-C 6 H4 -COOH = -O-CH 2 -CH2 -O-CO-C6H4 -CO- + 2H 2 O;
the polymer formed is polyethylene terephtalate (PETP).
c. P = number average degree of polymerisation, p = degree of conversion.
d. The molar mass of one unit is 192 (= 166 + 62 – 2·18), so 40.000/192 =

1/(1 – p) and p = 0.9952. Mw is 192·(1 + p)/(1 – p) = 79,800 g/mole, nearly twice

Mn .
Answers 9

e. Growing chains can close to rings; small acid or basic contamination can
terminate chain growth, etc.

2.29. With anionic polymerisation (“living polymers”) all chains start at the same
time and they grow at the same rate until the monomer is exhausted; the chains are,
therefore, equally long.

2.30. M n : ebulliometry (boiling point increase); measurement of end group

concentration (both for relatively low molar masses); osmometry for higher, but not
too high molar masses.

Mw: light scattering.

Mz: sedimentation in solution by ultracentrifugeing.
Further: see MT 2.2.3.

2.31. c (g/dl) t (sec) (t – to )to (t – to )/to c (dl/g)

0 100
0.05 106.25 0.0625 1.25
0.1 115 0.15 1.5
0.15 126.25 0.2625 1.75
0.2 140 0.40 2.0
In this oversimplified case the limit for c = 0 can easily be determined without
drawing a graph: [h ] = 1,0 dl/g.

2.32. [h] = k·Ma, so: 0.5 = k·(5·104 )a (1)

2 = k·(8·105 )a (2)
1 = k·(x)a (3)

From (1) and (2) it follows: 4 = 16a, so a = 0.5, and from (3) and (1): (x/5·104)0.5 = 2
so x = 200,000.

2.33. a is already known (a = 0.5); k follows e.g. from: 0.5 = k·(5·104 )0.5,
so k = 22.36 ·10–4.

2.34. a. Mn = 1/(0.98/20.000 + 0.02/200) = 6,711

Mw = 0.98·20,000 + 0.02·200 = 19,604

Mz = (0.98·20,0002 + 0.02·2002 )/19,604 = 19.996
— — —
so: very strong effect on Mn , small effect on Mw, and hardly any effect on Mz.

b. Here it is easier in kg/mole:

Mn = 1/(0.98/20 +0.02/2000) = 20,4 = 20,400 g/mol

Mw = 0.98·20 + 0.02·2000 =59,6 = 59,600 g/mol

Mz = (0.98·202 + 0.02·20002 )/59,6 = 138,490 = 1.384,900 g/mol
— — —
so: hardly any influence on Mn , considerable effect on Mw, very large effect on Mz.
10 Problems Polymers

2.35. We calculate ’/c, for convenience first in mm/(g/dl):

c (g/dl) ’ (mm) ’/c

0.12 6.50 54.2
0.18 9.94 55.2
0.25 14.05 56.2
0.33 18.8 57.0
0.45 26.6 59.1

Now we plot ’/c versus c. Extrapolation to c = 0 gives 52.8 mm/(g/dl).

1mm fluid pressure = r gh = 903 kg/m3 ·9.81 m/sec2 ·10 – 3 m = 46.7 Pa and

1 g/dl = 10 kg/m3 . It follows that: ’/c = 46.77 m2 /sec2 . This equals RT/Mn =
— —
8.314·298/Mn , so Mn = 52.97 kg/mole or 52,970 g/mole.

2.36. Measurements of osmotic pressure provide an absolute determination (without

calibration) of the number-average molar mass. This is independent of the type of
solvent; for each solvent the extrapolation to zero concentration results in the same

value: ’/c = RT/Mn . With viscosimetry the determination of M is not absolute;
dependent on the solvent and the temperature one finds a value for the intrinsic
viscosity, [h], which is not unique but which has to be calibrated.






0 0.1 0.2 0.3 0.4 0.5 c
Figure A.2. Figure to the answer of problem 2.35.

2.37. The increase of viscosity from a certain concentration becomes more

pronounced with increasing M (see Mark-Houwink), but the osmotic pressure is
inversely proportional to the molar mass.
— —
2.38. with p = 0.99, P n = 1/(1 – p)= 100, with p = 0.995, P n = 200. The mass of the
unit link is: (-CO-(CH2 )4 -CO-NH-(CH2 )6 -NH-) = 12·12 + 2·14 + 2·16 + 22·1 = 226
— —
g/mole. For A, therefore: Mn = 100·226 = 22,600 g/mole, for B: Mn = 200·226 =
Answers 11

45,200 mole. The weight-average degree of polymerisation is P w = (1 + p)/(1 – p) =
199 for A and 399 for B. So the weight average molar masses are 199·226 = 44,970
(for A) and 399·226 = 90,170 (for B). (end groups disregarded for convenience).
— —
2.39. wA = 0.3; wB = 0.7: (Mw)A = 44,970; ( Mw)B = 90,170;

Mn = 1/(0.3/22.600 + 0.7/45,200) = 34,770 g/mole (see 2.12),

Mw = 0.3·44,970 + 0.7·90,170 = 76,610 g/mole,

2.40. We can measure hrel = h /ho or h sp = (h – h o)/h o at a single concentration c

(from the run times to and t),and calculate hred = (h – ho)/h o c. hrel or hsp are in itself
already convenient data to rank polymers of the same kind after their molar mass. If
a quantitative estimation of [h] is desired, then these values can be corrected by the
known general dependence of hred on c.

2.41. The principle of Problem 2.40 is the basis of the k-value, used with PVC. A
measurement of the running time through a viscometer at a fixed concentration,
compared with that of the pure solvent, results, after a standard calculation, in the k-
value, which is, for that polymer, a simple index for the molecular mass. In
technological terms one knows that a PVC with k = 70 (high M) does not flow easily,
but it provides excellent end-use properties, e.g. in the long-term strength of pipes. In
injection moulding a better flow is required; then a PVC with k = 55 is more

2.42. The melt flow index is a useful indication of the molar mass, since it is a

reciprocal measure of the melt viscosity h . h depends very strongly on: h (:) Mw3,4,

(doubling of Mw results in a 10.6 times higher h !). This relation is valid for the
‘zero-shear’ viscosity; the melt index is measured at a shear stress where the non-
Newtonian behaviour, and thus the width of the molar mass distribution, is already
playing a part (see MT § 5.3.2). The melt index is a functional measure for the molar
mass, because for a producer of end products the processability is often of primary

2.43. From the results of the light scattering experiments it follows that A and B
— —
have the same Mw. The viscosity average, Mv , is for A higher than for B, and is thus
— —
closer to Mw. A has, therefore, a narrower molar mass distribution and a higher Mn .
The osmotic pressure ’ is therefore lower.

2.44. Large coils diffuse through the porous gel via the broadest channels, thus
following the shortest way, and they arrive as the first at the end. Smaller coils are
able to follow all side paths and have a longer travelling time.

The result of a GPC-measurement is a curve of wi against Mi from which Mw can be
calculated. Conversion to w i/Mi) against Mi and to wi ·Mi against Mi gives the
— —
possibility to calculate Mn and Mz as well.
12 Problems Polymers

2.45. – head-head and head-tail polymerisation in irregular sequences (exception).

– atactic instead of iso- or syndiotactic structure (example: PP).
– irregular branching (example: PE).
– with unsaturated chains: cis-, trans- and 1.2 configurations distributed at random
(examples: BR and IR).
– with copolymers: random distribution of building blocks (examples: SBR and

2.46. Isotactic: all R-groups are at the same side of the main chain; syndiotactic: in
alternating positions; atactic: randomly situated. All of this has to be considered in
3D space due to the tetraedric structure round a C-atom (see MT 2.3).

2.47. An atactic structure is in both cases not crystallisable. Atactic PP is because of

its glass-rubber transition temperature (T g = –15 °C) rubbery and technically of no
use. Isotactic PP is able to crystallise and can, therefore, be used in practice. For PS
atacticity is no objection; its properties as a glassy polymer are retained up to its Tg
(95 °C).

2.48. – by a stiff main chain, e.g. with rings (example: aromatic polyamides).
– by big side groups (example: PS).

2.49. – strong interactions: dipole forces, e.g. in PVC (-Cl) and PMMA
(-COOCH3 ).
also: hydrogen bridges, e.g. in PA and PU (NH....OC).
– weak interactions: dispersion forces, e.g. in PE and PP.

2.50. cross link density = number of connected units / total number of units. Another
measure is the molar mass between cross-links, Mc, The degree of cross-linking can
be determined experimentally from the swelling in a solvent or from the E-modulus.
vernettingsdichtheid = aantal verbonden eenheden / totaal aantal eenheden.

2.51. Chemical cross linking is the formation of primary chemical bonds between
chains, such as sulphur bridges (rubbers) or polystyrene bridges (UP), both between
unsaturated chains. In saturated chains cross linking can be brought about after
radical formation by irradiation or peroxides. Physical cross linking is based on
secondary bonds (e.g. domains in thermoplastic elastomers) or entanglements.

3. Glassy state and glass-rubber transition

3.1. Similarities: Both are not able to crystallise because of their irregular molecular
structure. Both will, therefore, reach the glassy condition upon cooling down from
the liquid phase.
Answers 13

Differences: When a low-molecular glass is heated it will, at its T g , immediately

change into a liquid (though less abrupt as at a real melting point). A polymeric
glass, however, first becomes rubbery as a result of chain entanglements. At a higher
temperature (Tv) it passes very gradually into the fluid state.

3.2. All three quantities (V, H and S) show an upward jump at Tm upon heating
(apart from “pathological” liquids such as water as regards V). At T g there is,
however, no jump, so DV = 0, DH = 0 and DS = 0, but a bend, or a jump in their first
derivatives. For V, H and S this means, respectively, jumps in the coefficient of
expansion a, in the specific heat c and in the compressibility k.

V, H, S V, H, S

Tg T Tm T
Figure A.3. Figure to the answer of problem 3.2.

3.3. Because of the many entanglements a long chain is still fixed at a number of
points, so that only parts of the chain (of the order of 50 to 100 links) are free to
move. At further temperature increase these entanglements are also loosened, so that
the polymer is able to flow (Tv region).

3.4. – Large side groups increase the distance between the chains and thus decrease
the interaction forces and, consequently, the Tg ; example: polyhexene-1 (see also
Question 3.6).
– Plasticisers, added in a sufficient quantity, also remove the chains from each
other. Example: PVC with di-octylphtalate, where the polar attraction forces are
strongly reduced and Tg may be lowered by more than 100 °C.

3.5. A thermodynamically defined transition is a transition between two equilibrium

states. The glassy state is, however, never in thermodynamic equilibrium; it is
subjected to volume-retardation (rearrangement of chain segments) up to unimagin-
ably long times. The glass-rubber transition should, therefore, be considered as a
freezing-in phenomenon, governed by kinetics.

3.6. The side group in polyhexene-1 is a straight and flexible C4 chain; the one in
PMP is, however, branched and rigid. Both side groups act as spacers between the
chains, thus tending to a lower Tg , but with PMP this effect is overcompensated by
the chain stiffening as a result of the rigid side group.
14 Problems Polymers

3.7. The difference between the Tg ’s of PP and PE is caused by the much greater
chain stiffness of PP as a result of the hindered rotation round the C-C bonds in the
main chain; the CH3 -groups are in the way.
In PVC rotation is hindered in a similar way (the Cl atoms are of about the same
size as the methyl groups). The Cl atoms, bound to the main chain, are strong
electrical dipoles and they exert much larger attraction forces on each other than the
methyl groups in PP, which are subjected to the much weaker dispersion forces only.
The same reasoning applies to the series BR, IR, CR. Here again we see an
increasing Tg . Now, however, less methyl- and Cl-groups are present, namely one on
four C-atoms instead of one on two. There is no effect of the double bond: this is
present in all three polymers in the same way.

3.8. PMMA: CH3 p.methylacrylate: H

| |
(110 °C) -CH2 -C- - - - (5 °C) -CH2 -C- - - -
| |
| |

This difference in Tg is a result of the enormous effect of the methyl group at the
main chain on the chain flexibility.
poly(n-butylmethacrylate): CH3
(25 °C) |
-CH2 -C- - - -
O-CH2-CH2 -CH2-CH3
poly(n-octylmethacrylate): CH3
(–20 °C) |
-CH2 -C- - - -
O-CH2-CH2 -CH2-CH2 -CH2 -CH2 -CH2-CH3

Increase of the length of the (flexible) side group increases the distance between the
chains and lowers Tg . Moreover, the polarity of the side group decreases, with the
same effect on Tg .

3.9. – The polymer may show a secondary transition in the glassy region as a result
of small movements of parts of the chain or side groups (example: PC).
– A blend of two incompatible polymers also shows two transitions: the Tg’s of
both components (example: TPS, toughened PS, PS + BR).
– A block copolymer tends to segregation; then there are two phases present with
each their own Tg (example: SBS).
Answers 15

3.10. SBR is a random copolymer of styrene and butadiene, and it has, therefore, a
Tg between that of poly-butadiene (–90 °C) and polystyrene (+95 °C). Its position
depends on the S/B ratio, which is chosen in such a way that the polymer can
function optimally as a technical rubber for tyres.

SBS is a three-block copolymer of styrene and butadiene. Incompatibility of PS and

poly-butadiene (BR) causes segregation of the chain parts: a two-phase system
results in which PS and BR each have their own individuality and their own Tg . The
fact that the highest Tg is lower than the Tg of normal PS, is a result of the relatively
low molar mass of the PS chain ends. In general the effect of M on T g is
insignificant: Tg = T•g - K/ M, but when K is, e.g. 105 and M = 10,000, Tg can be 10°

3.11. With copolymers it is of importance whether a one-phase or a two-phase

structure is formed (random- or block-copolymers). In a similar way we see with
polymer blends two (extreme) cases: real blends of wholly compatible components,
and dispersions of incompatible ones. In the first case, which seldom occurs (e.g. PS
with PPE) we find a single Tg ; in the second case two Tg ’s, namely those of the two

3.12. Since the glass-rubber transition is characterised by large chain parts becoming
mobile (e.g. 50 monomer units), we can from a single Tg only conclude that the
blend is homogeneous on that scale; at a smaller scale (of a few links) a two-phase
system may still be present.

3.13. The glass transition is mainly governed by chain interactions and chain
mobility. Cross-linking will hardly influence the interactions, but the mobility is
hindered by the cross-links. Tg will, therefore, increase with increasing network

3.14. When a chain with M = 200,000 g/mole is linked to other chains at four points,
the average molar mass between cross-links, Mc, amounts to 40,000. The mass of
one unit is 4¥12 + 6¥1 = 54 g/mole; so the number of units between cross-links is
about 740. At the glass-rubber transition no whole chains obtain free mobility, as a
result of the entanglements, but chain parts of 30 to 100 monomer units. The
chemical cross-links, therefore, hardly contribute to the restriction in chain mobility;
the increase in Tg will, therefore, be negligible.

3.15. Polyamides (nylons) can absorb water up to a few percent, which acts as a
plasticiser (the distance between the chain segments is increased), so that Tg is
lowered. Tg is, therefore, dependent on the environment (dry air, moist air or water).
The extremes may differ by several tens of degrees.
16 Problems Polymers

4. Crystalline polymers
4.1. A polymer which is in principle crystallisable because of its regular chains, may
crystallise too slowly to develop, under normal circumstances, a detectable degree of
crystallinity. In such a case the kinetics play the major role.

4.2. PC and PPE (Tg 150 and 215 °C) would have melting points of, respectively:
3/2 (150 + 273) – 273 = 361 °C(asymmetric chain) and
2(215 + 273) – 273 = 703 °C (symmetric chain).
In both cases processing above Tm would lead to strong degradation.

Poly-isobutylene (PIB) is a very useful rubber because of its very low gas
permeability. Co-polymerised with small amounts of isoprene (to enable
vulcanisation with sulphur) to butyl rubber (IIR), it is the ideal rubber for inner
tubes. If PIB would crystallise, it could not be used as a technical rubber! The same
holds for the rubbers BR and IR.

4.3. Even with a high cis content the rate of crystallisation of BR and IR is so low,
that generally speaking, no crystallisation occurs. However, upon stretching a few
hundred percent, micro-crystallites are being formed spontaneously. These form
extra cross-links between the chains, thus increasing the strength of the vulcanisate
enormously. When the stress is released these crystallites disappear, so that the
rubbery behaviour is completely restored. The same holds for butyl rubber, which
has, apart from a few isoprene segments, a regular chain structure.

4.4. Tm = DH/DS, in which DH is the increase of enthalpy upon melting, and DS the
entropy increase. DH is related to the chain interactions, DS to the chain flexibility.

4.5. Tm would be about 3/2 times Tg , so (3/2)¥(273+85) – 273 = 264 °C. Processing
of semi-crystalline PVC would only be possible above T m ; however, the polymer
cannot withstand such high temperatures.

4.6. n M Tm (°C) Tm (K) 1000/M 1000/Tm

8 114 –56.8 216.4 8.77 4.62
10 142 –29.7 243.5 7.04 4.11
15 212 10.0 283.2 4.72 3.53
20 282 36.8 310.0 3.55 3.23
32 450 69.7 342.9 2.22 2.92

After plotting 1000/Tm against 1000/M, and extrapolating the slightly curved line to
1000/M = 0, it appears that the first part of the curve can be represented by:
1000/Tm = 2.44 + 220/M.
So the melting point of PE with infinite chain length is 1000/2.44 = 410 K or 137 °C.
For Tm = 409 we find: M =37,000. With this low molar mass Tm appears to be only 1
Answers 17

°C lower than the maximum; so for practical grades the differences are tenths of a

5 1000/T m



3 32

0 1 2 3 4 5 6 7 8 9 10
Figure A.4. Figure to the answer of problem 4.6

4.7. The coiled structure with entanglements does not allow complete crystallisation
to occur. From very dilute solutions perfect crystals can be formed, in particular
when the chain molecules are lying fully stretched and oriented. A technical example
is the high-strength PE fibre: “Dyneema”.

4.8. The difference in T m must be due to the previous history. The first sample has
been quenched from the melt, and thus it is crystallised at a lower temperature, e.g.
130 °C. The crystallites are, therefore, smaller (more nuclei at a lower temperature)
and less perfect than in the second sample, which was cooled more slowly and which
crystallised at, e.g. 145 °C.

4.9. The lower the temperature of crystallisation, the lower the melting point of the
best crystal will be, but also: the more imperfect crystallites are formed. The melting
range of the sample with Tm = 161 °C is, therefore, broader.

4.10. PBTP has about the same chain structure as PETP, but has in the main chain
four CH2- groups instead of two with PETP. The chain is , therefore, more flexible,
which results in a lower melting point.
PBTP crystallises considerably more rapidly than PETP. In an injection moulding
process, where short cycle times are required, and thus rapid crystallisation during
the rapid cooling in the mould, PBTP may, therefore, be preferred. PETP would stay
amorphous, and, therefore, not usable above its Tg (69 °C).

4.11. Nucleating agents can be added to the polymer, which enhance the rate of
18 Problems Polymers

crystallisation considerably. It is also possible to take the amorphous product from

the mould and heat it for some time in an oven at some temperature above its Tg , so
that it crystallises to form an acceptable end product.

4.12. Chain orientation enhances the rate of crystallisation, because every volume
element in which a few chain parts lie parallel forms a nucleus. This is the case in
fibre spinning, film stretching and bottle blowing.

4.13. In a (stereo-specific) rubber spontaneous crystallisation occurs under strain;

here also small regions in which chain parts lie parallel, act as nuclei.These nuclei,
which improve the strength considerably, do, however, not grow out into a
continuous phase; they disappear upon stress release.

4.14. Nuclei with a radius r smaller than rk are disrupted by the thermal motion;
when r > rk the nuclei are stable and can grow. rk is determined by the competition
between the formation of new interface (interfacial energy has to be supplied) and
the production of crystallisation heat. In the formula for r k , (Tm –T) is in the
denominator, which indicates that with stronger super-cooling a nucleus can grow
more easily.

4.15. The ratio of the crystallisation rates is 98/23 = 4.26, the ratio of the molar
masses is 143/87 = 1.64. The ratio of the melt viscosities (which govern the crystal
growth) amounts to 1.643,4: 1 = 5.38. Apparently the rate of crystallisation is, next to
the rate of nucleation, governed by the melt viscosity.

4.16. The Tg values of most of the polyamides are between 50 °C (PA-6 en PA -6.6)
up to 82 °C (PA-4.6) (in ambient atmosphere a few tens of degrees lower due to
moisture absorption). They, therefore, already lose their properties at some T
increase. Crystallisation is, therefore, essential for technical applications
Amorphous polyamides should have a higher Tg (e.g. Gelon A 100 of General
Electric with Tg = 130 °C) by the presence of aromatic rings in the main chain, and,
moreover, an irregular chain structure (to prevent crystallisation), by the introduction
of side groups.

4.17. With increasing n the -CO-NH- groups in the main chain are more and more
“diluted” in the -CH2- sequences; their effect decreases and Tg and T m approach to
the values of PE. The “zig-zag” effect is a matter of parity: the formation of
hydrogen bridges is only possible if the groups are in the proper position.

4.18. - For a non-crystallisable polymer:

a. and b. equal
c. solidifies earlier
d. volume-retardation
Answers 19

- For a crystallisable polymer:

a. undercooling
b. melting range
c. more undercooling
d. post-crystallisation.


d b

T2 Tg T1 T

Figure A.5. Figure to the first part of the answer to problem 4.18.


d b

T2 Tm T1 T

Figure A.6. Figure to the second part of the answer to problem 4.18.

4.19. a. Tg will certainly lie between 20 °C and 80 °C; probably between 30 en 70°.
b. PETP has an asymmetric chain, so Tm = (3/2) ¥ T g , i.e. between (3/2)(273+30)-
273 = 182 °C en (3/2)(273+70)-273 = 241 °C.
c. Above Tg crystallisation can occur in the rising branch of the curve for
crystallisation rate vs temperature.
d. At T m the rate of nucleus formation is zero; at Tg (and at Tm) the rate of crystal
growth is zero ; between these two a maximum occurs (see MT § 4.3.2).
e. See Figure A.7.
f. See Figure A.8.
g. In the rapidly cooled mould of an injection moulding machine crystallisation
proceeds too slowly and needs help from nucleating agents or annealing above Tg
after moulding. See also problems 4.10, 4.11 and 4.12 .
20 Problems Polymers

V d

d c T ∞C

0 50 100 150 200 250 300

Figure A.7. Figure to the answer of problem 4.19e.

Log E

after C

after A

T ∞C

0 50 100 150 200 250 300

Figure A.8. Figure to the answer of problem 4.19f.

4.20. The volume of the unit cell is 0.740 ¥ 0.493 ¥ 0.253 nm3 = 9.230¥10-29 m3.
When n CH2- groups are present in such a cell, then the mass within the cell is m = n
¥ 14 g/mole, so n¥14/ 6¥10 23 grammes = 14n/6¥10 26 kg. The density is: r =
(14n·1029)/(6·1026·9.23) = 253·n kg/m3 . n must be an integer and (must be slightly
higher than the density of a near-linear PE (960 kg/m3 ). Only n = 4 meets these

4.21. 1 cm3 of the polymer has a mass of r grammes. It contains j cm3 of

crystalline polymer with a mass of j¥ rc grammes and (1–j) cm3 amorphous with a
mass of (1–j) ¥ ra gr. The total mass is j = (r – ra)/(rc – r a)..
1 gramme of the polymer has a volume of 1/r cm3 and contains (1/r ) ¥ j ¥ rc
grammes crystalline matter, so the crystalline mass fraction is y = j·(r c/r) = (rc/r)·
(r – ra)/(rc – r a).

4.22. The Avrami equation can be transformed into:

log [–ln(1 – DV/DV•)] = log c + a·log t.

Answers 21

We can now construct the following table:

t V DV l = DV/DV• 1–l –ln(1 – l)

0 120.00 0 0 1
500 119.16 0.84 0.084 0.916 0.0877
1000 114.97 5.03 0.503 0.497 0.6992
1500 110.94 9.06 0.906 0.094 2.364
2000 110.04 9.96 0.996 0.004 5.521
2500 110.00 10.00 1.000 0

Plotting – ln(1 – l) against t, both on a log scale, indeed results in a straight line with
slope 3, so a = 3. c can e.g. be determined at t = 1000: c = 7·10–10 .

log(–ln(1 – l))




2.5 3 3.5 log t

Figure A.9. Figure to the answer of problem 4.22.

b. a =3 can mean: three-dimensional crystal growth with heterogeneous nucleation

of two-dimensional growth with homogeneous nucleation.
c. At 155 °C the exponent a will not change very much. c will, however, be
considerably smaller as a result of much slower crystallisation.
d. The volume of the amorphous polymer is 120 mm 3; so the amorphous density is
r a = 100/120 g/cm3. The density of the crystallised sample is r = 100/110 g/cm3.
With r c = 0.937 g/cm3 the crystalline volume fraction can be calculated:
jc = (r – ra)/(rc – r a) = 0,73.

4.23. From jc= (r – ra)/(rc – ra) it follows for LDPE: jc = 0.38

and for HDPE: jc= 0,76.

4.24. The cause can be found in the morphology of the polymer. The question is: is
the amorphous phase continuous or the crystalline phase or both? This is further
dealt with in Chapter 9.

4.25. A semi-crystalline polymer is a two-phase system; it consists of a crystalline

22 Problems Polymers

phase and an amorphous phase, which, in most cases, have a different refractive
index. Light is, therefore, deflected at each interface, which results in strong light
When, by exception, the refractive indices are equal (as in PMP), no light scattering
occurs and the polymer is transparent.
When the crystallites are small enough (small in comparison with the wave length of
the light), only minor scattering occurs. This can be achieved with nucleating agents,
but also with chain orientation.

4.26. In a super-strong PE fibre the chains lie nearly completely stretched, and are
thus loaded under optimal conditions. They must, however, transmit the mechanical
stress on each other, which can only happen via transverse bonds. These are
relatively weak binding forces (London-Van der Waals bonds), so the number of
these bonds should be made as large as possible, which means very long chains.

4.27. For a semi-crystalline polymer the E-modulus shows between Tg and Tm (in
which region it is already lower than below Tg ), a rather strong decrease at increasing
T, whereas with amorphous polymers, which are used below Tg , the stiffness is not
much temperature dependent (apart from possible secondary transitions). The time
dependency, or the creep, shows a similar behaviour.
As a matter of fact, this does not hold for semi-crystalline polymers far below their
Tg . (strong example: PEEK). The word “far” is here important, since creep proceeds
up to very long times of loading, whereby Tg is lowered; the polymer can as it were
creep towards its Tg !

4.28. Lyotropic LCP’s are processed from a solution, thermotropic ones from the
melt. In both cases the flow pattern provides the necessary orientation of the stiff
Working with a solution is needed for polymers which above their melting point
would degrade (example: aromatic polyamide fibres such as Kevlar and Twaron).
For fibres the removal of the solvent is not too problematic. In e.g. injection
moulding applications solvents cannot be used; here thermotropic LCP’s have to be
used. Since these would degrade during processing, they are “diluted” by
copolymerisation (example: poly-hydroxy-benzoic acid - co - PETP)

4.29. The glass-rubber transition extends over a rather broad temperature region. Tg
is, therefore, dependent on the chosen method of measurement and on the definition
of the transition point. Moreover, Tg depends on the time scale of measurement.
Though the melting point, Tm , is a well-defined quantity for an ideal crystal, melting
also extends over a broad temperature region as a result of the differences in crystal
perfection. The highest melting point found for a polymer depends on the quality of
the “best” crystal, which depends on the regularity of the chain structure (compare
Answers 23

HDPE and LDPE) and on the rate at which the polymer has crystallised (so on the
crystallisation temperature).

4.30. All three transitions are shown by a semi-crystalline thermoplastic with a chain
length long enough to extend the rubbery region to above the melting point.

4.31. Only T g and T v are present in an amorphous thermoplastic and in a non-

vulcanised rubber. A semi-crystalline thermoplast with short chains only shows Tg
and Tm. In a (very exceptional) fully crystalline polymer Tg is not present

4.32. For vulcanised rubbers and for thermosets Tg is the only transition.

4.33. Between Tg and T m the polymer contains a crystalline and an amorphous,

rubbery phase. The rubbery phase is responsible for the high impact strength.

5. Rubbertoestand en vloeibare toestand

5.1. F is the free energy and is related to the Gibbs free enthalpy G, via G = U + pV
– TS = F + pV. When F is used instead of G the work exerted on the environment by
volume change is thus neglected. U is the internal energy, which increases when
work is done to overcome interaction forces. T is the temperature in Kelvin. S is the
entropy of the system, and is a measure for the probability of the state, or for the
number of possibilities (conformations) at which the state can be realised.

5.2. When the length l is increased the force needed is:

dF dU dS
K = dl =
dl – T dl

In the glassy state interaction forces have to be overcome, so dU/dl is important. The
entropy, however, hardly changes, since the chain conformations do not change at a
small deformation. The force is, therefore: K = dU/dl.
In the rubbery state, on the contrary, the chain interactions are not or hardly active:
they have, from T g , been overcome by the thermal movement. The entropy, S,
strongly depends on the deformation so the force is now given by: K = –T · dS/dl.

5.3. The entropy change upon elongation is a matter of conformations, so purely

topological, and is not influenced by temperature.

5.4. a. The molar mass of the isoprene unit (-CH2 -C.CH3 =CH-CH2 -) is 68 g/mole,
of a half cross-link: 4 ¥ 32 =128 g/mole. The mass fraction S is 0.005. When n
isoprene units are between crosslinks, then 128 = 0.005 ¥ n ¥ 68, so n = 376 and
Mc = 376 ¥ 68 = 25,600 g/mole.
b. Substitute: r = 910 kg/m3, R = 8.31 J/K.mole, T = 298 K, Mc = 25.6 kg/mole: E =
24 Problems Polymers

264.000 Pa.
c. E appears to be more than 5 times higher. Strain-induced crystallisation cannot be
an explanation, for E is measured at very small strains. Deviation from ideal
rubber-elastic behaviour gives a very small effect only. The explanation is the
existence of physical entanglements, which are much more numerous than the S-
bridges. (without vulcanisation also a value of nearly 1.5 MPa! is found!)
d. Entropy-elasticity dominates the behaviour: T¥dS/dl is the important parameter. S
is a matter of order/disorder: the number of possible conformations, The
difference between conformation possibilities in strained and unstrained condition
is greater as the freely moving chain part is longer; therefore Mc is in the

5.5. a. Mc = 3 rRT/E = 3 ¥ 910 ¥ 8.31 ¥ 293 / 2.2¥10 6 = 3.02 kg/mole = 3020

g/mole. This amounts to 3020/54 = 56 monomer units. The use of proper units is
very important ! In this case, i.e. for non-vulcanised rubber: Mc = molar mass
between physical, temporary entanglements. Not the whole chain!
b. E(100) = E(20) ¥ 373/293 = 2.8 MPa, but: entanglements are loosened; there is a
gradual transition to flow, so M c increases and E decreases. The three deviations
from ideal rubberelastic behaviour, mentioned in MT 5.1, are here, for non-
vulcanised rubber, of much less importance.

5.6. a., b., c. See Figure A.10.

d. With light cross-linking Tg hardly increases, since the chemical cross-links are at
much greater distances from each other than the physical entanglements. With
tighter networks this is no longer the case, so that the chain mobility is hindered
by the presence of the many cross-links and Tg is increased.
e. The same reasoning as above: In the formula E = 3rRT/Mc, the distance between
cross-links, (expressed as molar mass M c) is, with light vulcanisation, mainly
governed by the physical entanglements. With higher degrees of cross-linking the
distance between chemical bridges plays an increasing role, so that E increases.
f. For an ideal network T is in the numerator of the formula for E, so that E is
proportional to the absolute temperature. The log E - T curve thus shows a
positive slope (not a straight line because of the log-scale but slightly curved
upward). In reality this simple picture is often disturbed by deviations from ideal
rubber-elastic behaviour.
Answers 25

log E



a. T
Figure A.10. Figure to the answer of problem 5.6.

5.7. When the rubber is heated from 253 K to 298 K, the E-modulus becomes
298/253 = 1.18 times higher. The wall stress in the balloon will increase by 18% and
the diameter decreases to 20/1.18 = 17 cm. Gradually the air in the balloon will be
heated up, and the pressure will increase proportionally to the absolute temperature.
Eventually the old diameter of 20 cm will be reached.

5.8. First: Also in the rubbery condition flow will occur and in the fluid state the
melt also shows elastic properties.
Secondly: the transition from rubber to flow is strongly dependent on the molar
mass: this transition is governed by the unravelling of chain entanglements, which
are, with long chains, much more numerous than with short ones.
Thirdly: the longer one takes the time to disentangle the coils, the easier this is
performed. The time scale plays, therefore, also a large role. At short times a
polymer may behave as a rubber, whereas it flows at longer times. (think e.g. of
“silly puttee”).

5.9. a. en b.: see Figure A.11. A twice as high molar mass gives a 23.4 = 10.6 times
higher zero-shear viscosity, in which the relevant average is Mw is (h (:) M 3‚4 w )
The whole curve is thus shifted upward by a factor 10 (one unit on the log scale).
Broadening of the molar mass distribution shifts the curve to the left; the zero-
shear viscosity does not change.
c. the “power-law” holds for the right-hand part of the curve, which often resembles
. .
a straight line. From t = k·(g)n it follows, together with t = g·h: h(:)t(1–1/n). When
n = 12 h(:) t–1, so on the log-log scale the slope is –1 (45° to the left).
d. In injection moulding shear stresses and shear rates are high, so that a broad

distribution is beneficial for easy mould-filling. However, the high Mz may be
responsible for high melt-elasticity resulting in frozen-in orientations and
26 Problems Polymers


103 b.


0,1 1 10 100 1000 t Pa

Figure A.11. Figure to the answer of problem 5.9.

With rotation moulding t is very small; a narrow distribution may then be

preferred, also because a higher Mn favours the impact strength etc. With film
extrusion a broad distribution is advantageous for making very thin films and for
the elastic effects (Mz) required in shrink films; a narrow distribution results in
better tear strength and impact resistance (Mn ).

5.10. In a plot of log h vs. log t the curves are shifted vertically upon a change in M
or in T. (see MT Figure 5.12). In such a plot the melt-index can also be indicated
easily, since the m.i. is measured at a fixed stress.

5.11. a. The melt flow index is measured at a shear stress which may strongly differ
from the one present in a processing operation. In particular with injection
moulding the latter is much higher than the former.
b. By a higher value of the shear stress, so a greater force on the plunger, better
agreement can be reached, e.g. instead of 2.16 kgf: 5 or 10 or even 21.6 kgf.
c. At e.g.10 kgf the ratio of the melt flow indices is higher than 10 /2.16 ; this ratio
increases with increasing width of the molar mass distribution (non-Newtonian
behaviour or ‘shear-thinning’).

5.12. An increasing melt-flow index signifies decreasing molar mass; with polymers
— —
made in the same process both Mw and Mn are lower. Since the impact strength is
— —
governed by Mn , it decreases with increasing m.f.i. With a narrower MMD Mn is,
however, higher than what would be expected on the basis of the m.f.i; therefore the
impact strength is also higher.

5.13. A higher m.f.i. means a lower molar mass and a lower melt viscosity. The rate
of crystallisation will then be higher, and a grade with a higher m.f.i. will crystallise
better and reach a higher degree of crystallinity, thus a higher stiffness. For the

grades with a narrower MMD one would expect, at the same m.f.i., a lower Mw, so a
lower zero-shear viscosity (see MT, Figure 5.14), which would result in a higher
Answers 27

crystallinity and a higher stiffness. It is, however, the other way round. Possibly in
the crystal growth process higher shear stresses and -strains are involved than under
m.f.i. conditions; possibly other, unknown causes play a part.

5.14. Polyethylene crystallises much more rapidly than polypropylene; under usual
cooling conditions PE will always reach its maximum possible crystallinity,
irrespective of its chain length, contrary to PP. An exception is formed by the ultra-
high-molecular PE (UHMWPE), in which the chains are exceptionally long (M > 106
g/mole), the viscosity exceptionally high and the crystallinity (also the density)
slightly lower. However: super- strong PE -fibres are made from the longest possible
chains, but the crystallinity is extremely high as a result of the near-perfect alignment
of the chains.

5.15. a. “die-swell”. i.e. recoil of the elastic deformation of the melt after it leaves
the die of an extruder;
b. “melt fracture”: the elastic strain may be so high that the material breaks; the
product is distorted;
c. crystallisation: with a crystallisable polymer (e.g. a stereospecific rubber) the
elastic strain may reach such high values that spontaneous orientation-induced
crystallisation occurs, causing considerable trouble in the process (see MT Figure
d. when the elastic deformations (chain orientations) are frozen in during cooling, an
anisotropic end-product results, which shows different stiffness and strength in
different directions. Also the refractive index may be dependent on the direction
(bi-refrigence, which is of importance with e.g. compact discs).

6. Visco-elasticity
6.1. A Maxwell element shows an instantaneous elastic deformation and thereafter
unlimited flow. For polymers in the solid condition the latter is not realistic. For a
fluid polymer it is more relevant; moreover, the instantaneous elastic deformation is
in accordance with the real behaviour: when the stress is released a polymer fluid
shows an instantaneous recoil.

6.2. a. eo = s/E =104/105 = 0.1 = 10%.

b. e = eo + s·t/h = 0.1 + 104·500/107 = 0.6 = 60%.
c. s relaxes between t = 500 and t = 800 sec with s = soexp(–t/t), in which so =
104 , t = 300 sec and t = h/E = 100 sec, so s = 104·e–3 = 500 Pa (e3 is close to 20).
d. See Figure A.12.
e. - Non-linearity. - spectrum instead of single relaxation time
28 Problems Polymers

s kPa

0 200 400 600 800 t sec




0 200 400 600 800 t sec
Figure A.12. Figure to the answer of problem 6.2

6.3. Evidently a fluid polymer cannot be considered; in the model the deformation
approaches to a limit. For a solid polymer the model seems more appropriate, though
is represents neither a spontaneous elastic deformation nor permanent flow.
Therefore a combination of a Kelvin-Voigt element with a spring and with a dashpot
in series is, in principle, more appropriate.

6.4. Cross-section of the bar = A = 10–2·10–3 = 10–5 m2 ; the stress s = K/A = 600 N/
10–5 m2 = 60·106 N/m2 = 60 MPa.
The strain is Dl (mm)/l (mm) = Dl(mm)/100, so for 1 mm strain is e = 0.01.
a. The model shows instantaneous strain and retarded strain, and it can, therefore, in
principle be represented by a spring E1 in series with a Kelvin-Voigt element with
E2 and h .
b. From the instantaneous strain (2 mm) it follows: e = 0.02. So E 1 is: E 1 = s /eo =
60/0.02 = 3000 MPa.
The limit of the strain is 8 mm, so e = 0.08, of which the Kelvin-Voigt element
accounts for 0.08 - 0.02 = 0.06 , so in e = s/E1 + (s/E2 )·(1 – exp(–t/t)) is s /E2 =
0.06, so E2 = s/0.06 = 60 MPa/0.06 = 1000 MPa.
At last the retardation time t: at t = 250 sec e = 7,5 mm = 0.075. Substitution:
Answers 29

0.075 = 0.02 + 0.06·(1 – exp(–250/t) results in: 0.055 = 0.06 · (1 – exp(–250/t));

t = 100 sec = h/E2 , so h = t·E2 = 100·109 = 1011 Pa·s.
c. When the stress is released after 250 sec the strain on E 1 disappears
instantaneously; so the total strain decreases from 7.5 to 5.5 mm. Thereafter the
strain decreases as: e = 5.5 * exp(–t/t), where t starts at 250 sec.
d. Non-linearity and spectrum rather than a single retardation time.

0 100 200 300 400 500 600 sec
Figure A.13. Figure to the answer of problem 6.4.

6.5. a. eo = s/E2 = (6·106 )/(3·109) = 0.002 = 0.2% (only the spring in series).
b. e(3·106 ) = eo + e1 = 0.002 + (s /E1)·(1 – e–t/t), (t = h /E1 = 106 sec) = 0.002 +
(6·106 )/(3·108)·(1 – e–3) = 0.002 + 0.02·0.95 = 0.021 = 2.1%
c. E2 recoils: e = 2.1 – 0.2 = 1.9%.
d. The Kelvin element recoils with: e = 0.019·exp(–3·106 /106) = 0.019·0.05 =


0 0 1 2 3 4 5 6 ¥ 10 6 sec

Figure A.14. Figure to the answer of problem 6.5.

30 Problems Polymers

6.6. The course of the strain can be described by superimposing the effects of the
two stresses; the first one is s, applied at t = 0, and continued after t1; the second one
is –s, applied at t = t1 . We find then from t = t1:
e = (s/E2 ) + (s/E1 )·(1 – exp(–t/t)) – (s/E2 ) – (s/E1 )·(1 – exp(–(t – t1)/t) =
= (s/E1 )·{exp(–(t – t1 )/t ) – exp(–t/t)} =
= (s/E1 )·{1 – exp(–t1/t )}·{exp(–(t – t1)/t )} =
= e(t1) · exp(–(t – t1)/t ),
in which e(t1) is the strain at t1 immediately after the release of the stress.
Without the superposition principle we find the same result : After taking away the
stress the spring E2 is unloaded and we keep the deformed Kelvin-Voigt element
with a strain: e = (s/E1 )·(1 – exp(–t1 /t )) = e(t1).
The spring E 1 now pulls back the Kelvin-Voigt element from t = t1 with a stress s,
proportional to the remaining deformation e: s = E 1 ·e(t – t1).
The rate of deformation is given by the behaviour of the dashpot: de /dt = –s/h
(negative, since the deformation decreases).
So s = E 1·e(t – t1) = –h·de /dt, from which: e = e(t1)·exp(–(t – t1)/t) with, again, t =
h/E1 .

6.7. The strain after removal of the stress approaches tot 0, and there is also an
elastic strain of 1 cm (retraction from 5 to 4 cm strain), A possible model is,
therefore, a Kelvin-Voigt element (with E 1 and h parallel), in series with a spring E2.
From the immediate jump in the strain after stress release (s = 104 N/10 –4 m2 = 108
Pa), it follows: e2 = 1/20 = 0.05 = s/E2 = 108 /E2 , so E2 = 20·108 Pa = 2000 MPa.
Now the Kelvin-Voigt element: from Dl = 4·exp(–2t) it follows t = 21 hour = 1800 sec
= h/E1 . The strain during the first 5400 sec can thus be represented by:
e = e2 + (s/E1 )·(1 – exp(–t/t )) = 0,05 + (108 /E1 )·(1 – exp(–3)) = 0.5, from which
(with e 3 = 20) follows: E1 = 475 MPa. With t = 1800 sec this results in: h = 1800·
4.75·108 = 855·109 Pa·s. The model is shown in the figure.

6.8. Now the strain, after stress release, approaches to 0.5 cm, which is a permanent
strain e3 = 0.5/20 = 0.025 as a result of the dashpot in series in the model. The
viscosity, h2, of this dashpot is given by: e3 = s·t/h2 or 0.025 = 108 ·5400/h 2 , so (h2 =
21.6·1012 Pa·s.
The elastic deformation is now 0.5 cm or e 1 = 0.025,, so the spring in series has a
stiffness s/ e1 = 108 /0.025 = 40·108 Pa = 4000 MPa.
For the Kelvin-Voigt element remains:
e2(5400 sec) = (4 cm) = 0.2 = 108/E1 ·(1 - exp(–5400/1800)) = 0.95·108 /E1 , so E 1 =
475·106 = 475 MPa. Since t = 1800 sec, h1 = 1800·475·106 = 855·109 Pa·s.
The model is, therefore, as shown in Figure A.16.
Answers 31

E1 = E1 =
h = 855 GPa·s h1 = 855 GPa·s
0.475 GPa 0.475 GPa

E2 = 2 GPa E2 = 4 GPa

h2 = 21.600 GPa·s

Figure A.15. Figure to the answer of Figure A.16. Figure to the answer of
problem 6.7. problem 6.8.

6.9. Let the deformation of the spring E be e1 and of the dashpot: e2. Their sum, e =
e1 + e 2 is equal to v·t, in which v is the constant rate of strain, de/dt. The stress is
given by s = E·e1 = h·(de2/dt).
With (de 2 /dt) = v – (de1/dt) we find: v = (E/h)·e1 + de 1 /dt, from which, after solving,
it follows that: e1 = v·t·(1 – exp(–t/t)) with t = h/E.
So thhe s - e diagram (see Figure A.17) looks as follows: s = E·v·t·(1 – exp(–t/t) =
h·v·(1 – exp(–t(t)).


Figure A.17. Figure to the answer of problem 6.9.

6.10. Again: e = v·t. Now the stress on the spring is s1 = E· e = E·v·t, and on the
dashpot: s 2 = h ·v. The total stress is s = v ·(E · t + h ) = (e/t)·(E·t + h ) =
e·(E + h/t) = e·E + h·v.
32 Problems Polymers


Figuur A.18. Figure to the answer of problem 6.10.

6.11. Apparently the stress relaxation proceeds in two phases, each with a stress
decrease of 1 MPa at a strain e = 1. It seems logical to think of a parallel arrangement
of two Maxwell elements, both with a spring constant E = 1 MPa, but with relaxation
times which differ by a factor 10,000 . Inspection of the stress values (with s =
so·exp(–t/t)) easily results in: t 1 = 1 sec, t2 = 10,000 sec. The viscosities are then: ( h
= t·E): 106 and 1010 Pa·s.

6.12. a. E 1 is the storage modulus, and indicates how much reversible energy is
stored in the material (purely elastic). E2 is the loss modulus; it is a measure for
the energy lost during deformation i.e. dissipated into heat; it thus signifies the
contribution of the fluid nature of the material.
b. Vibration experiments, e.g. with a torsion pendulum, yield E1 from the vibration
time and E2 from the damping of the vibration (see MT 6.2).
s MPa

0,01 0,1 1 10 102 103 104 105 t sec
Figure A.19. Figure to the answer of problem 6.11.

c. The ratio of these two quantities, E2/E1 , is tan d, which is, therefore, a measure for
the position of the material in the visco-elastic region (d = 0: purely elastic; d = 2p
(90°): purely viscous)
d. – heating under repeated stress cycles (e.g. with fatigue).
– damping of shock waves, e.g. with rubber-modified polymers,
– use of a polymer for damping of vibrations,
– heating of an automotive tyres,
Answers 33

– friction of a tyre on the road (see MT 7.3)

6.13. a. a. The curve p is typical for an amorphous, not cross-linked polymer; q for a
cross-linked polymer,
b. For both curves the maximum lies at the glass-rubber transition temperature Tg,
c. For p the polymer gradually shifts into the liquid condition; d increases up to p2
and tan d to infinity,
d. In a network (q) the loose chain ends (not held between two cross-links) do not
contribute to the network elasticity, but they behave more or less as a fluid.
Therefore, with increasing T, tan d approaches to a finite value, dependent on the
network structure.

7. Mechanical properties
7.1. The cross section of the bar is 10–2·10–3 = 10–5 m2 . The stress is s = 600/10–5 =
60·106 Pa = 60 MPa. The elongation is e = 2/100 = 0.02. The modulus of elasticity is
E = s/ e = 60/0.02 = 3000 MPa. = 3 GPa.

7.2. A value for E of about 3 GPa is normal for an amorphous polymer in the glassy
state, unless below Tg a strong secondary transition occurs such as with PC, so that
the E-modulus at ambient temperature is significantly lower.

7.3. With semi-crystalline polymers we should always carefully distinguish between

the behaviour below Tg and above Tg . Below Tg (such as with PEEK) the crystalline
fraction, which is somewhat stiffer than the amorphous glass, dominates, so that E is
somewhat higher. Above Tg , such as with PE and PP, the amorphous fraction, which
is in the rubbery condition, is responsible for a significantly lower E.

7.4. – By chain orientation, such as in fibres,

– by “reinforcing” fillers, such as hard particles or short fibres.

7.5. Tighten both bars at equal length in a clamping-screw, bring them in a bending
vibration and listen to the pitch. The highest tone is given by the stiffest bar, which is
the bar made of pure PP. Musical people may be able to estimate the interval
between the two tones; if this is, e.g. a whole-tone distance (12% in the frequency),
we know that the ratio of their E-moduli is about (1/1.122) = 0.8. With the Kerner
formula (Chapter 9) we can now estimate the rubber content.

7.6. Fibres are, as a result of the spinning process, molecularly oriented, and they
have, therefore, a 2 to 3 times higher stiffness than the non-oriented polymer (e.g.
polyamide and polyester textile fibres). With the highest attainable orientation, such
as in aromatic polyamides (Twaron and Kevlar), and in the PE-fibre (Dyneema) the
stiffness can be a hundred times higher than the one in the unoriented condition!
34 Problems Polymers

7.7. The chain stiffness influences the height of the glass-rubber transition
temperature (and of the melting point), but not the stiffness of the polymer below Tg
(in the glassy state). Extremely stiff chains show the effect of the formation of LCP's
(liquid-crystalline polymers), by which very high stiffness is reached, but only in the
direction of the orientation.

7.8. In principle a shifting rule could be applied, but the volume retardation
(physical ageing) is responsible for a gradual increase of the resistance against creep,
so that the rate of creep gradually decreases. Creep curves at e.g. 20 °C thus proceed
considerably less steep, and the real creep is therefore (fortunately!) much less than
would follow from the T-log t shift (see MT 7.2.4).

7.9. See Figures A.20, A.21 and A.22.

a. could possibly apply to a semi-crystalline polymer with a Tg not too far above
ambient temperature, so that E (below Tg ) is still high; at longer times of loading
the polymer A “creeps gradually towards Tg" (example: PETP).
b. could apply to an amorphous glassy polymer; E has its normal high value (about 3
GPa), while T g is high enough to be not approached at longer times of loading
(example: PMMA).
s MPa s MPa s MPa
10–2 1 1 10–2
20 102 20 10–2 102 20
104 hours 104 hours

10 10 10

104 hours
e e e
0 0 0
0 1 2% 0 1 0 1 2%
Figure A.20 — 22. Figures to the answer of problem 7.9.

c. The low E-modulus could indicate a semi-crystalline polymer above Tg such as

PP; the strong creep of a crystalline - rubber two-phase system is also in
accordance with this assumption.

7.10. The Kohlrausch formula is well suitable to represent the creep at small stresses
and strains. D o is then the compliance at t = 0, and is a measure of the immediate
elastic deformation. The formula, however, fails when the creep behaviour is non-
linear; this is, in general, the case with stresses occurring in practice.

7.11. a. After cooling to below Tg the volume of a glassy polymer continuously

decreases gradually (see MT 3.3.): "volume-retardation".
b. The resistance against creep increases strongly by volume retardation.
Answers 35

c. See Figure A.23.

s s

e e
without with physical ageing

Figure A.23. Figure to the answer of 7.11 c.

7.12. Noryl (homogeneous blend of PPE and PS) is an amorphous polymer in the
glassy state. It shows relatively little creep; the isochrones lie close to each other.
With PBTP the distance between isochrones is much greater: there is more creep.
This is related to the semi-crystalline structure of PBTP (contrary to PPE/PS).
Though the glass-rubber transition temperature of PBTP, Tg, is about 65 °C, this
transition is already apparent at lower temperatures and longer loading times. The
polymer belongs, therefore, to the category described in MT 4.5, Figure 4.20.

7.13. a. Can be read off directly from the 102 hour isochrone at 8 MPa: e = 1.33 %.
MPa 1
5 8
6 16 s 10–2 hours
14 102
5 10
8 12
4 5
4 8
3 5

3 4 6
2 4
1 2

100 80 60 40 20 ∞C 1 2 3 4%

Figure A.24. Figure to the answer of problem 7.13a.

b. At 70 °C we choose the axis half-way between 60 °C en 80 °C; for 10 hours an

isochrone half-way between 1 and 100 hours, and we find at 5 MPa: e = 1.94%.
36 Problems Polymers

c. For the short-time modulus we take the slope of the 10–2 hour isochrone, and we
read off at 20°, e.g. at 0.2% : 3.4 MPa. It follows then that E = 3.4 / 0.002 = 1700
d. With the same isochrone we find at 80o and 0.2%: 1.3 MPa, so E = 1.3/0.002 =
650 MPa.
e. 1 year = 8766 hours, so close to the 104 hours isochrone. With e = 2% we find at
20 °C about 8 MPa as an allowable stress.
f. At 100 °C and 1000 hours (half-way between the 102 and 104 isochrones), we
find for 3% about 3.7 MPa.

7.14. Plotting times to failure (log t, horizontal) versus applied stress (log s, vertical,
different scale), results in the following figure:

20 ∞C
40 ∞C
60 ∞C
80 ∞C
100 ∞C

50 110 ∞C

time to failure
10–2 10–1 1 10 102 103 104 105 hours
Figure A.25. Figure to the answer of problem 7.14.

From this figure, safe wall stresses at various temperatures and loading times can be
determined by extrapolation. We extrapolate the curves found at higher T to lower
T’s. Here a ten years life (87,660 hours) at 80 °C is required. Extrapolation of the
curve for 80 °C results in an estimated stress to break of 28 MPa. To remain at the
safe side, we stay somewhat below this value, so that 20 MPa looks a reasonably safe
wall stress for this purpose.
Answers 37

7.15. a. First damage can occur by "crazes" (see MT 7.4.2)

b. To avoid "crazes" completely, we have to stay below a "critical strain", which, for
each polymer is about a material constant. To remain within this limit the use of
creep isochrones is required..

7.16. First: A well defined notch forces the sample to break at that place because of
the stress concentration, and not at an accidental scratch or contamination. The
results are, therefore, better reproducible.
Secondly: At the location of the notch the rate of deformation is considerably higher
than in the remainder of the sample. The higher rate of strain has a similar effect as a
temperature decrease, so that one measures, as it were, at a lower T, with a much
greater chance of brittle fracture (which is important for practical use) (see MT

7.17. a. Improvement of impact strength can be reached by adding a dispersed

rubbery phase.
b. The rubbery phase introduces a glass-rubber transition at the Tg of the rubber,
and, therefore, a damping peak which is able to absorb shock energy (analogous
to some secondary transitions in the glassy region, such as with PC). Moreover
we can think of the ‘neutralisation’ of a just initiated crack in a dispersed rubber
c. At the Tg of the added rubber the E-modulus of the blend necessarily decreases; in
the temperature region of use we, therefore, lose in stiffness.
d. Examples: TPS (toughened polystyrene) or HIPS (high-impact PS) or SB
(styrene-butadiene); also PP copolymer, a PP with "tails" of ethylene-propylene
copolymer, which form a rubber phase, PVC with CPE (chlorinated PE), etc. etc.

7.18. The impact strength of a polymer is strongly dependent on the molar mass, in

particular on Mn (number average, number of chain ends)

7.19. a. In a Shore-D hardness test the penetration of the pin is measured during
loading; the load is thus prescribed as well as the resulting deformation (see MT
b. The time scales of both measurements may differ considerably; moreover, with a
hardness test we may enter into the non-linear region.
c. A hardness test is, with a simple apparatus, much easier than a measurement of
the E-modulus, which requires the measurement of a stress-strain diagram and the
determination of its initial slope.

7.20. The higher the temperature above Tg , the more we are removed from the
damping peak, so the lower is tan d.

7.21. With increasing T, tan d approaches to a limiting value, which is governed by

38 Problems Polymers

the network perfection, in particular by the number and length of loose chain ends,
which as it were represent a fluid component.

8. Further properties
8.1. With an amorphous thermoplastic the brittleness temperature is about Tg, unless
a secondary transition temperature Tsec occurs in the glassy region; in that case the
brittleness temperature may be in the neighbourhood of Tsec.
- with a semi-crystalline polymer the brittleness temperature is about Tg.
- with a rubber-modified thermoplast the brittleness temperature is about the Tg of
the dispersed rubber.

8.2. With an amorphous thermoplast the polymer softens over a rather short
temperature interval from the glassy to the rubbery state. With a semi-crystalline
polymer a certain amount of softening takes place at Tg; with further T- increase the
stiffness drops very gradually up to the melting point Tm.


ISO-A from table 60 55 60 95 90 115 45
ISO-A estimated from 40 25 50 80 90 115 65
Vicat B from table 180 110 90 100 230 165 70
Vicat B estimated from 150 130 140 90 240 180 110

The agreement is reasonable, though not perfect. Differences can be accounted to:
- small differences between the samples taken,
- differences in the time scales of measurements,
- non-linear behaviour (in particular with the Vicat test).

8.4. a. The contradiction is a result of the slope of the log E - T curves (see MT
b. The ISO-A test is carried out at a higher E-modulus level than the Vicat test, thus
giving more information for applications where a higher E is required, i.e. in
loaded condition.

8.5. The blending of short glass fibres results in an increase of the E-modulus to e.g.
its threefold over the whole temperature region. If the slope of the log E - T curve is
small. such as with semi-crystalline polymers between Tg and T m ,, a vertical shift
causes a considerable horizontal shift (see MT 8.1.2.), so a strong increase of the
softening points, which are of importance in applications at higher temperatures.
Answers 39

8.6. The high coefficient of expansion can cause problems with shape stability and
dimension tolerances in precision products. Moreover the high expansion causes
more shrinkage upon cooling after a shaping process (see e.g. MT 10.3.2).

8.7. Advantages:
- various applications in which, e.g. a handle remains cool ; also in foams for
thermal insulation it is important that the cell walls do not conduct too much heat.
- With repeated changes of load the dissipated heat cannot easily flow away, so that
the temperature of the object increases and it loses its good mechanical properties.
- In a processing operation the heating-up and, thereafter, the cooling-down, are
troublesome operations because of the low thermal conduction. In heating-up use
can be made of friction or dielectric effects etc. (see MT 10.1.1), but with cooling
the low heat conduction mainly governs the cycle time.

8.8. a. (see MT 8.1.4): l = lp + lg + lr + lk ( polymer conduction + gas conduction

+ radiation + gas convection).
b. Larger cells : more convection; larger l k , so l increases.
c. Lower density: lower l p , but higher l r (radiation passes through cell walls),
therefore a minimum in l.
d. l g depends on the type of gas (heavier gas: lower lg).
e. The lowest possible minimum is l g , the coefficient of conduction of the gas in

8.9. Firstly: Though a high softening temperature is attractive, we should realise that
the polymer has to undergo processing operations, which are carried out, for an
amorphous polymer, well above T g , and for a crystalline polymer, above Tm . At
these temperatures the polymer should be sufficiently chemically stable (for i.a. that
reason PPE is being blended with PS).
Secondly: For long-term applications not only the softening point is of importance,
but also the degradation behaviour of the polymer; it may quite well happen that the
polymer suffers unallowable degradation considerably below its softening T (see MT

8.10. a. The term “stress corrosion” signifies damage (crack formation) under
simultaneous presence of a mechanical stress and a certain environment, which both
in itself are not fatal.
b. The word “corrosion” suggests that a chemical reaction is involved, but in several
types of stress corrosion no chemical degradation occurs, but it is caused by a
lowering of the interfacial energy (such as with PE with a surface active agent).
Another example of “physical” stress corrosion is the effect of e.g.
tetrachloromethane on PC.
40 Problems Polymers

c. In those cases when chains are actually broken we have to do with real stress
corrosion, such as the effect of ozone on a rubber vulcanisate.
d. see above.

8.11. – To avoid risks of electrostatic charging (sparks, explosions) anti-static

additives are sometimes added to a polymer; also carbon black may be used (see MT
– Polymers are sometimes used to shield electromagnetic fields; for this purpose
they can be filled with metal fibres.
– In extreme cases conjugated dienes with an adequate doting can reach an electric
conductance similar to that of metals.

8.12. The electrostatic charging is mainly controlled by the specific resistance of the
material (sometimes also by the surface resistance). Decrease of the tendency to
build up electrostatic charges can, therefore, be reached by lowering the specific
resistance or the surface resistance, e.g. by adding carbon black or short metal fibres,
or by using anti-static agents. The time scale is important because each application
presents its own requirements as regards the dissipation of the charge, e.g. the
formation of sparks or the attraction of dust (see MT 8.2.1).

8.13. Advantages:
- The low tan d of PE at very high frequencies enabled the application of radar in
the 1930’s.
- PE cannot be heated in a high-frequency electric field for, e.g., welding films.
Special “tricks” are then needed, such as coating the film with a thin layer of a
polymer with a higher tan d.

8.14. a. With precision optics dimensional tolerances are extremely narrow, so that
even the smallest amounts of recovery are disastrous for the performance.
b. Advantages for spectacle glasses are: reduced weight and higher impact resistance.
Disadvantages are the lower resistance against scratching (which can be improved by
a coating), and, on the long term, the dimensional stability.

8.15. One of the main problems with a compact disc is the birefringence; the laser
beam is split-up into two components, thus mutilating the signal. It is, therefore,
important to reduce the birefringence; this can be done by minimising the chain
orientations by choosing a PC with a low M. This is, moreover, beneficial for the
ease of processing (very tiny details!), but is at the cost of the impact strength, which
is, fortunately, for this type of application already high enough.

8.16. IIR (butyl rubber) is compared with other rubbers (SBR, NR) and then has a
Answers 41

20 to 30 lower permeability (inner tubes!). PE and PP are much more permeable than
most of the other plastics. These phenomena can be attributed to the fact that the rate
of diffusion in the rubbery phase (large free volume) is considerably higher than in
the glassy-, and, in particular, in the crystalline phase. PE en PP contain, next to a
crystalline phase, a continuous amorphous rubbery phase.

8.17. In the first place the rate of diffusion of LDPE is higher, as a consequence of
the higher amorphous fraction. Secondly, for the same reason, LDPE has a higher
solubility (the solubility in the crystalline phase is very low).

9. Polymeric compounds and composites

9.1. The molar masses are reflected in V1 en V2 (the molecular volumes). Increase of
these values causes the first two (negative) terms to approach more to zero, so that a
negative value of DGm becomes less probable.

9.2. c is related to the enthalpy of blending, and is positive when only dispersion
forces are present between the chains, so that, in view of the small contribution of the
entropy of blending (the first two terms), DGm will be, in most cases, positive. With
strong interactions c becomes negative, and, at the same time, DGm..

9.3. For j1 = j2 = 0.5 and V2 = V1 the formula reads::

DGm/V = kT[–(ln 2)/V1 + c/4Vs]. This is negative when c/4Vs < (ln 2)/V 1 of: c <
2.77(Vs/V1 ).

9.4. At very small values of j 1 , ln j 1 is so deeply negative that DGm is negative,

independent of the values of V 1 , V2, c etc. The same holds for j2 . The curve for DGm
as a function of f thus proceeds, at its extreme values, steeply downward,
independent of how its further course looks like, e.g. as indicated in the figure.


Figure A.26. Figure to the answer of problem 9.4.

9.5. When there is a UCST (upper critical solution temperature), then above that
temperature the components are fully mixable. When cooling down to below this
UCST a region of segregation can be reached, in particular when the volume
fractions do not differ too much (see MT Figure 9.3, lower part "upside down")
42 Problems Polymers

9.6. Now about an LCST (lower critical solution temperature, see MT Figure 9.3). If
we blend below this temperature, we can obtain a homogeneous mixture. Subsequent
heating to above the LCST can result in segregation; the rapid cooling in an injection
moulding machine does not allow to undo this.

9.7. For thermodynamic reasons mixability may be all right, but the last step in a
mixing process, when the droplets are sufficiently small, is a diffusion process; this
needs a considerable time, which is not always available in a blending process (in
this way a "multiphase" blend could be formed, see MT Figure 9.4.d)

9.8. With very short "blocks", such as in the random-copolymer SBR, there is only a
very small difference in entropy between the segregated and the homogeneous
condition: no or hardly any change of entropy upon segregation. As the block length
increases, this difference, however, increases, so that (G decreases (see MT 9.1.4).
Segregation is, therefore, less complete in multi-block copolymers (such as
polyethers- polyesters).

9.9. On the one hand in a flow field shear stresses t are exerted on a droplet which
cause a deformation into an ellipsoid; on the other hand the surface area of the
droplet is increased by this deformation, so that the interfacial energy s/R increases.
When the size of the droplet decreases the first effect becomes smaller in comparison
to the second one, which results in an equilibrium at which no droplets are broken-up
(expressed in the capillary number Ca = t/( s/R), (see MT 9.1.5).

9.10. The droplets may have been broken down to their minimum diameter, but in
the flow field they meet, they collide, and they can coagulate into bigger ones. The
blending process is, therefore, an ongoing competition between break-up of droplets
and coagulation.

9.11. The first process is: the deformation of droplets into an ellipsoid, until the
droplet breaks up. The second process is: A droplet is, at high speed, extended into a
thread, which is subject to instabilities (according to Rayleigh and Tomatika), and
which, therefore, breaks up into a row of droplets (see MT 9.1.5).

9.12. Series- and parallel arrangements provide a first approximation of the

properties; the properties which follow from these arrangements, are widely
different. Combinations of series- and parallel arrangements (Takayanagi) are able to
provide good approximations, but they require a large number of parameters (see MT

9.13. If we apply Kerner’s formula to this PE, with jcr = 0.552, for a continuous
crystalline phase, in which an amorphous phase is dispersed, we find j = 0.448, A =
1.14 (with n = 0.33), Eo = 6000 MPa, E 1 = 1 MPa (both estimated provisionally), so
Answers 43

a = 1/6000.
1 + 0.448·1.14·(1/6000 – 1)/(1/6000 + 1.14)
E = 6000· 1 – 0.448·(1/6000 – 1)/(1/6000 + 1.14)
= 2378 MPa

With the inversed morphology, so crystalline dispersed in amorphous, we have: j =

0.552, Eo = 1 MPa, E1 = 6000 MPa, so a = 6000, A = 1.5, and, according to Kerner:
1 + 0.552·1.5·(6000 – 1)/(6000 + 1.5)
E = 1· 1 – 0‚552·(6000 – 1)/(6000 + 1.5) = 4.08 MPa

With Nielsen’s formula we find: E = [0.552·60000,2 + 0.448·10,2]5 = 598.5 MPa.

Apparently the result of Nielsen’s formula is closest to the observed E-modulus.
With a few trials the values of Ecr = 6400 MPa and E am = 1.5 MPa render the best
agreement, viz. E = 667 MPa, compared to the observed value of 660 MPa.
Conclusion: In PE the crystalline phase and the amorphous phase are both
continuous and "interwoven".

9.14. a. 150S + 2000B + 150S = (with S = 104, B = 54) = 15,600 + 208,000 +

15,600 = 139,200 g/mole; the weight fraction styrene w(PS) = 31,200/139,200 =
b. The mass of a PS end chain is 15,600/6.1023 g = 2.60·10–20 g. The volume of a PS
domain = (p/6)·25 3 ·10–21 cm 3 = 8.18·10 – 1 8 cm 3 , its mass V ·1.05 =
8.59·10–18 g. Divide: the number of PS chains in a domain is 330
c. Main cause: T g is lower for PS domains because of its low M, compared to
"normal" PS; the difference may be as high as 10 °C. (MT 3.4.3). Therefore at
Tg -10 °C softening of the PS domains can already occur.
d. First we calculate the volume fraction j of PS: 1 gramme SBS contains 0.224 g
PS, so 0.224/1.05 = 0.213 cm 3 PS, and 0.776 g BR or 0.776/0.91 = 0.853 cm3 BR.
Total volume is 1.066 cm3 , of which 0.213 cm3 PS, so j = 0.213/1.066 = 0.20.
Substituting this, with a = 1000 and A = 1.5 (rubbermatrix), into the Kerner
equation, results in: E/Eo = 1.62 so E = 4.9 MPa.
The PS domains not only act as cross-links, but they also increase the stiffness.

9.15. When a foam contains closed cells, we can apply Kerner’s formula with a = 0,
so: E/Eo = (1 – j)/( 1 + j/A).
With j = 0.5, Eo = 3200 MPa and A = 1.14 (for n = 0.33) we find: E = 1112 MPa.
An open cell foam is a co-continuous structure, sot that Nielsen’s formula can be
applied: En = j 1 E n1 + j 2 E n2 with n = 0.2, j1 = 0.5, E1 = 3200 and E 2 = 0 (gas!); this
results in: E = (12 )5 ·3200 = 100 MPa.
The conclusion is that the first foam has closed cells, the third one is completely
open-celled and the second one contains partly open, partly closed cells.

9.16. For a "heavy" foam the gas fraction j is small. The Kerner formula for a foam
44 Problems Polymers

with closed cells: E/Eo = (1 – j)/(1 + j/A) can then be approximated by: E/Eo = 1 –
j – j /A = 1 – (1 + 1/A)·j. For rubbers with A = 1.5 this becomes: E/Eo = 1 – 1.7·j,
for glassy polymers with A = 1.14: E/Eo = 1 – 1.9·j . When we use the
approximation to 1 – 2j or (1 – j)2 (always for small j), and we realise that 1 – j is
proportional to the foam density, then from this approximation it indeed follows, in
particular for hard thermoplastics, that: E (:) d2
The approximation seems somewhat rough, but it appears to fit reasonably.

9.17. With the introduction of particles in a polymer the stiffness increases, but the
tensile strength and the impact strength are hardly increased. It is, therefore, better,
to use the term: ‘stiffening additives’.

9.18. Carbon black increases the tensile strength of an SBR vulcanisate to its 10 to
20-fold (which would, otherwise, be very low), as well as its abrasion resistance.
Natural rubber can, because of its stereospecific (cis) chain structure crystallise
under strain, and, therefore, reach higher values of its tensile strength; for a good
abrasion resistance carbon black is also of importance with NR.

9.19. A continuous phase forms a "skeleton" in the blend. This is important, e.g. in
the following cases:
– a second phase with a considerably higher softening temperature; when the other
phase softens, the skeleton retains its stiffness.
– an electrically conducting polymer as a second phase provides the wanted
– an impermeable polymer as the second phase renders the blend impermeable only
if it forms a continuous network.

9.20. See Figure A.27.

log E

A M1 M2 B


T °C

0 100 200
Figure A.27. Figure to the answer of problem 9.20.

Apparently all four curves intersect at the same point, namely at the temperature at
which A and B have the same E-modulus. Each blend has the same modulus at this
temperature. This situation may occur when B is a glassy amorphous polymer and A
Answers 45

is semi-crystalline. The behaviour of Xenoy (PC + PBTP) resembles this picture.

9.21. See Figure A.28.

The four curves have about the same shape; they are shifted along the T-axis. C and
D are typically amorphous polymers. They are wholly miscible. The pattern
resembles the one of Noryl (PPE + PS).

log E

C M3 M4 D


0 100 200 T °C
Figure A.28. Figure to the answer of problem 9.21.

9.22. Apparently two opposing effects play a part: On the one hand the glass fibres
take over a part of the stress, thus partially releasing the stress on the polymer. On
the other hand, fibres cause, in particular at their ends, stress concentrations, which
can initiate crack formation.

9.23. E o = 3000 MPa, E1 = 2 MPa, so a = 1/1500, so a can safely be assumed as

zero in Kerner’s formula. From n = 0.33 it follows: A = 1,14. The value aimed at is
E = 2400 MPa. From Kerner it follows: 0.8 = (1 – j)/(1 + j/1.14), so j = 0.1175.
easy rule of thumb: the percentage of modulus reduction is about twice the volume
percentage of rubber, or also gas bubbles).
The mass fraction is: y = j·rr/rpr in which rpr is the density of the polymer - rubber
blend: y = j · r r/[(1 – j)·r p + j · r r] = 0.1175·0.9/(0.8825·1.4 + 0.1175·0.9) =
0.106/1.34 = 0.079 = 7.9 % w.

9.24. In the blend of the previous question we blend glass fibres. Now Eo = 2400
MPa and E1= 75,000 MPa, so a = 31.25. E should be increased up to 3000 MPa so
E/Eo = 1.25. When we substitute this into the Kerner equation, again with A = 1.14
(not wholly correct since n has become somewhat higher because of the blending
with rubber), Kerner’s formula then becomes:
1.25 = [1 + j·(1.14·30.25/32.29)]/[1 – j·(30.25/32.39)], from which it follows:: j =
0.112 of 11.2 %. The mass fraction is now, calculated on the polymer-rubber blend
(r pr = 1.34): y = 0.112·2.4/(0.888·1.34 + 0.112·2.4) = 0.184 = 18.4 %w.

9.25. In a composite one aims at the highest possible stiffness E at the lowest
possible r , so at a high as possible value of E/r, and this is the square of the sound
46 Problems Polymers

velocity in the material.

9.26. The limit is reached when the glass fibres are so long that they can be
considered as a continuous phase. We have then a parallel arrangement of polymer
and glass, with j1 = 0.75, j2 = 0.25, E1 = 3 and E 2 = 75 GPa. The E-modulus of the
composite then follows from: E = 0.75 · 3 + 0.25 · 75 = 21 GPa.

9.27. The Einstein formula E = E o (1 + 2.5 j ) is valid if a is rather high and a is

rather small, and it gives, for the transverse direction: E = 3 (1 + 2.5 · 0.25) = 4.9
GPa. (The calculation with Kerner is better, but we now aim at a simple first

9.28. In an isotropic composite (chaotic distribution of fibre direction) E is given by:

E = 0.2 · Epar + 0.8 · Etransv. The first term follows from the answer to Question 19,
the second one from 20, and the result is: E = 0.2 · 21 + 0.8 · 4.9 = 8.1 GPa. This
value, therefore, approximates the highest attainable value with 25 vol% of glass

9.29. glass fibres rubber particles

a. Tg no change no change (there is an extra
transition, namely of the rubber
b. Vicat higher lower (shift effect E(T) curves, see
MT 8.1.2)
c. Tm no change no change
d. ISO-HDT higher (see b.) lower
e. E-modulus higher due to hard lower, due to soft inclusions
f. tensile strength slightly higher lower
g. heat conduction higher due to glass fibres about the same
h. impact strength sometimes better, considerably higher (see MT 9.3)
sometimes worse

10. Processing techniques

10.1. a. Heating from the outside (metal or infrared); controlled by heat conduction
b. Friction (between granules, controlled by the friction coefficient µ), or in the
liquid, (controlled by the viscosity h);
c. High-frequency dielectric heating, controlled by the dielectric loss factor tan d;
d. Ultrasonic sound, controlled by the mechanical loss factor tan d.
Answers 47

10.2. Casting is, of course, possible when the material is fluid enough. This is the
case with some thermosets (casting resins), which, together with a second
component, can be cast into a mould and then cured at a higher temperature.
With thermoplastics casting is, in general, not possible because of the high melt
viscosity, except when polymerisation has still to take place; the monomer or pre-
polymer is polymerised in the mould. This is, e.g. possible with PMMA and PA-6
(from methyl-methacrylate and caprolactam, respectively).

10.3. With semi-crystalline polymers the rubbery region is partly or wholly masked
by the crystalline phase. Only very high-molecular grades types have part of the
rubbery region left above their Tm. Processing techniques in which the rubber-elastic
behaviour is essential, such as calendering and vacuum forming, are therefore less
suitable for crystalline polymers, unless they have a very high molar mass.

10.4. When moulding thermoplastics the mould has to be heated, and, after shaping.
cooled down to enable the taking-out of the product. This requires repeated changes
in temperature of large blocks of metal. With thermosets and rubbers the temperature
can remain high after shaping to enable curing or vulcanisation.

10.5. a. Two methods may be applied:

– injection moulding,
– forming from an extruded sheet.
Both can lead to economic production. With the popular brown or white coffee-
cup the second method (two steps) has, apparently, won the competition.
Transparent PS cups are mostly injection-moulded.
b. The bottom of the injection-moulded cup shows the point of injection. Moreover,
after sheet forming the wall thickness is less uniform than after injection-
c. An injection-moulded cup will gradually soften upon heating, whereas a vacuum-
formed one first contracts into a flat sheet.
d. The choice between the possible processing methods is i.a. governed by the
economy of production; this is strongly related to the production size (see MT

10.6. a. At a certain injection pressure p, a minimum temperature T is needed to

reduce the viscosity sufficiently to enable complete mould filling. A lower T requires
a higher pressure.
b. The mould B requires a higher pressure to be filled; so it has a higher flow
resistance (narrower) or a longer flow path (l/d).
c. In B the melt flows less easily than in A (higher viscosity); at first sight we could
suppose that B has a lower melt flow index (m.f.i.)
d. If, however, the m.f.i. of B appears to be higher than that of A, then A has a much
48 Problems Polymers

broader molar mass distribution; it is stronger non-Newtonian, so that it flows

better at the high shear rates in the machine, despite of its lower m.f.i.
e. At further increase of the pressure the mould halves may be pushed apart, because
the clamping force is insufficient ("flash").
f. A too high temperature may lead to degradation of the polymer.
g. A too low T involves the risk of incomplete or non-homogeneous melting of the

10.7. a. When the pressure in the mould is 800 bar (1 bar = 105 N/m2 ), the material
exerts on the area of 100 cm2 a force of 800 ¥ 105 ¥ 10-2 N = 800 kN ("80
tonnes"). This is thus the minimum required clamping force.
b. In the mould a volume is present of 10¥10¥0.3 = 30 cm3. At atmospheric pressure
and 220 °C this would mean 30/1.34 = 22.4 grammes. At 800 bar the specific
volume is, however, a fraction 800 ¥ 7.5 ¥10-5 = 0.06 or 6% smaller, the mass is
6% greater, or 22.4 ¥1.06 = 23.7 g. For the end product, at 1 bar and 20 °C this
would mean a volume of 23.7 ¥ 1.05 = 24.9 cm3 or 24.9 / 30 = 0.83 times the
volume of the mould. This means a volume shrinkage of 17% or a linear
shrinkage of about 5%. The dimensions would be about 9.5 cm ¥ 9.5 cm ¥ 2.85
d. After filling of the mould extra space becomes available by cooling and
solidification along the walls; maintaining the pressure until the polymer at the
injection gate is also solidified will then enable several percents extra supply

10.8. PE, PP and PA are semi-crystalline polymers; melting and solidification go

accompanied by a (though gradual) volume jump. PS, PVC and PC are amorphous
thermoplastics; upon solidification they show no volume jump, but only a bend in
the V-T relation.

10.9. The outer side cools and solidifies first; the inner part remains fluid the
longest; its shrinkage in the final stage of solidification is constrained. Its density
will, therefore, be lower than at the outer side and tensile stresses are being built up,
even with the risk of void formation.

10.10. Powder (blend of resin and curing agent) is deposited on the metal via a fluid
bed or electrostatic. Upon heating it flows and, at the same time, it is cured. The
course of the viscosity in an analogous situation is given in MT Figure 10.8. Now,
however, too slow curing results in dripping of the fluid, while a too rapid cure
stagnates the flow so that the surface is not smooth.

10.11. Between the rolls of a calendar enormously high pressures are built up,
needed to transport and shape the material at high speed. These pressures cause
Answers 49

bending of the rolls, even though they are very heavily constructed. The result is that
the sheet is too thick in the middle. By crossing the axles of the rolls this can be
largely compensated (see MT 10.4.1).

10.12. With calendering and also with vacuum forming the polymer must be in the
rubbery condition; in the first case the sheet must be taken off from the last roll
under a certain stress, which it can withstand in the rubbery phase only. With
vacuum forming the heated sheet is placed on the mould; this is only possible when
it possesses enough coherence and handability. The word ‘molten’ is therefore not
relevant in these cases.
In other processes such as blow-moulding and film extrusion some rubber-elastic
response also plays a part, though considerably less than in the examples mentioned

10.13. First: a broad molar mass distribution helps to lower the apparent viscosity at
high shear stresses and shear rates, so that thinner films can be produced at the same

Secondly: a broader MMD leads to a higher value of Mz, the average which is
mainly responsible for the elastic behaviour of the melt, and thus for the recoil after
deformation and solidification. These frozen-in rubber-elastic deformations cause the
behaviour of shrink-films.

10.14. In rotation moulding the shear stresses and shear rates are extremely low; the
polymer particles flow together into a homogeneous wall under the influence of
gravity and surface stresses only. The relevant viscosity is, therefore, the
"Newtonian" viscosity, without the complication of ‘shear-thinning’.

So the width of the MMD is not of primary importance: only the weight average, Mw
counts. A consequence is, that for a given m.f.i. the distribution should be narrow
(see MT 5.4).
— —
At the same time, with a narrow distribution Mn is higher since it closer to Mw, This

is important for the resistance against stress corrosion, since this is related to Mn
(good for oil tanks).

10.15. Melt fracture results from a too large rubber-elastic deformation. This is
governed by the polymer behaviour and by the rate of deformation. The strain rate
must be lowered, which may be accomplished by a lower, and therefore unattractive,
rate of production. The time scale of the elastic deformation can also be increased in
another way, namely by making the conical channel to the die more pointed; the
same strain is then reached after a longer time. This goes at the cost of extra pressure
and power, since the resistance of the channel is increased.

10.16. When a pipe is extruded, the melt has to flow round the ‘spider’ at which the
50 Problems Polymers

torpedo is suspended, and, thereafter, flow together perfectly. The risk is an

imperfect welding, causing weaker spots at the weld lines or ‘spider lines’.
Also with injection moulding weld lines occur frequently. The best known example
is the eye for the handle of a pail, which is the spot at which the stresses are highest
during use!

10.17. Adiabatic: no exchange of heat with the environment. For an extruder this
means that no heat is supplied from the outside, but the energy for melting is wholly
supplied by the internal friction. This is ideal for homogeneous heating of the
material, without the trouble of low heat conduction. This can be realised with larger
diameters. In most cases the heating-up is a combination of wall heating and internal

10.18. Along the screw of an extruder both transport flow and transverse flow occur.
The latter provides a circulation flow, which promotes mixing and homogeneity.

10.19. a. Q is the throughput in volume per unit time.

p is the pressure in front of the die,
N is the rotational speed of the screw.
h is the viscosity of the melt,
a is a geometrical constant, related to the screw geometry.
b ditto,
c is a geometrical constant related to the die dimensions.
b. The maximum throughput is 80 dm3/hour; this is reached when there is no die at
the end of the extruder (drag flow only). The maximum pressure (for a given
viscosity of the polymer) is 400 bar (extrapolated); this is reached when the die is
c. When the rotation rate is halved, the screw characteristic shifts down in a parallel
way, and it then intersects the vertical axis at 40 dm3 /hour. Both the throughput
and the pressure are then halved.
d. When the die resistance is doubled, the slope of the line through the origin is
halved. From the graph it can then be deduced: the rate of throughput, Q is
lowered from 80 to 48 dm3 /hour, the pressure p increases from 100 to 160 bar.
e. When the viscosity becomes three times higher, both slopes are three times
smaller; the rate of throughput, Q, remains 60 dm3 /hour (this also follows from Q
= N ¥ a ¥ c/(b + c), in which h and p do not appear), but the pressure p becomes 3
times higher, so 300 bar.
f. As a result of the non-Newtonian behaviour both expressions for the pressure
flows (b¥p/h and c¥p/h) are no longer valid. The curve for the die is now curled
upward since the apparent viscosity decreases with increasing shear stress. Also
the shape of the screw characteristic changes.
Answers 51

Q (dm3/h)

70 (e)


50 (e)


(c) (b)

p (ba
0 100 200 300 400
Figure A.29. Figure to the answer of problem 10.19 f.

10.20. a. Internal stresses are physically present stresses; frozen-in stresses are
frozen-in chain orientations. The former result from uneven cooling from the
melt: solidification first takes place at the cooled outer wall, while the inner parts
solidify much more slowly, thus creating internal tensile stresses. The stress is,
averaged over the cross-section, zero, but there is a stress distribution, composed
of compression- and tensile stresses.
Frozen-in stresses originate from the rubber-elastic behaviour of the melt: the
rubber- elastic deformations (chain orientations) are frozen-in upon cooling and
remain present as ‘latent’ stresses.
b. Internal stresses (cooling- or shrinkage stresses) become apparent when from a
ring, cut from a pipe, a small segment is removed; the ring will tend to close
because the tensile stresses are mainly at the inside. Frozen-in stresses can be
detected by heating a piece of a pipe to above Tg or Tm; the rubber-elastic stresses
then get their chance to deform the sample to shorter and thicker dimensions.
c. Shrinkage stresses contribute to the total stress on a volume element in a pressure
pipe. In the pipe a circumferential stress is present, on which, at the inner side, the
shrinkage stress is superimposed. This may limit the life-time of the pipe.
Frozen-in orientations are present in the length direction of the pipe and may
cause anisotropy: in the transverse- (circumferential-) direction the material is
weaker, and this direction is just important for a pressure pipe! Moreover, at
higher temperatures, the pipe may deform.
d. With thick-walled pipes cooling proceeds more slowly so that the shrinkage
52 Problems Polymers

differences and the internal stresses become greater. Thin-walled pipes are
extruded at a higher rate; the deformation of the melt takes place at a shorter time-
scale, resulting in a higher rubber-elastic deformation. Moreover, the chain
orientation is more readily frozen-in because of the quicker cooling.

10.21. a. PETP crystallises only slowly, and it stays amorphous in a normal

injection-moulding process, after rapid cooling of the melt.
b. PETP has a Tg of about 70 °C and a Tm of 260 °C. If the pre-form would be semi-
crystalline, then it could only be blown-up at a temperature above Tm. At such a T
the viscosity is so low (condensation polymer, so not a very high M!), that
blowing is not possible. Amorphous PETP can, at some distance above its T g ,
easily be deformed in the rubbery condition.
c. During the blowing process the bottle is simultaneously stretched in the length
direction so that a bi-axial orientation is reached, which is useful for the
d. An amorphous bottle would soften completely already above 70 °C, so it would
be hardly usable.
e. Chain orientation, also bi-axial, promotes rapid formation of nuclei, which is
necessary to allow the polymer to crystallise sufficiently at a rather short time
f. Because of the very large number of nuclei formed upon chain orientation, the
crystallites remain small enough to avoid appreciable light scattering.
g. During spinning of polyester fibres such a high uni-axial orientation is brought
away that rapid crystallisation occurs. This is, of course, essential for a good

10.22. a. Higher m.f.i.: lower M, so, in general, less elastic behaviour of the melt
and a lower die-swell.
b. Larger ratio of m.f.i. at 10 and at 2.16 kg indicates a more strongly pronounced
non-Newtonian behaviour, so a broader MMD with relatively more long chains

(higher Mz), resulting in stronger elastic response of the melt and a higher die-
c. At a higher rate of throughput the rate of deformation of the melt is higher,
resulting in more elastic response and a higher die-swell.
d. At a higher temperature the relaxation times are shorter, so that there is less
chance of chain orientation when the melt leaves the die.
e. Most of the rubber-elastic strain is brought about in the conical channel leading to
the die (elongational flow).The longer the melt stays in the (straight) end channel,
the more the orientations are able to relax, and the lower the die-swell will be.

10.23. PVC is an amorphous thermoplastic; during the heating-up along the screw
of an extruder the viscosity is, after passing Tg , still very high and it decreases only
Answers 53

slowly. Compression of the mass must, therefore, take place over a considerable part
of the screw length. PA, on the contrary, is slightly above its Tm already a relatively
low-viscous fluid (also since its molar mass is not very high). The compression zone
may, therefore, be short.

10.24. Structural foam (or integral foam) is a foam where the walls are solid.
– Though its production, e.g. in an injection moulding machine, requires special
tricks, the injection pressures and the mould clamping forces required are much
lower than with a solid product.
– The surface is, compared with a normal foam, smooth.
– The product is considerably lighter than a solid product (e.g. half).
– The solid skin provides extra bending- and buckling resistance.

10.25. a. blow extrusion,

b. casting of monomer (caprolactam) and in-situ polymerisation,
c. sheet extrusion, followed by vacuum forming, or rather SPPF slightly below Tm.
d. injection moulding,
e. calendering and cutting,
f. centrifugal casting from a paste, followed by vulcanisation,
g. calendering,
h. compounding, extrusion, moulding, vulcanisation,
i. film extrusion (balloon), cutting, welding,
j. pre-foaming of pentane containing beads, final foaming in the mould,
k. moulding, vulcanisation,
l. moulding and vulcanisation, or: injection moulding and vulcanisation,
m. sheet extrusion or calendering, thereafter sheet forming (vacuum or with extra
n. casting from monomer, in-situ polymerisation,
o. wire-coating extrusion round a copper wire,
p. extrusion,
q. impregnating of glass wires with a blend of resin and curing agent, winding
round a mandrel, curing,
r. rotation moulding from powder,
s. moulding of a powdery compound, curing,
t. sheet extrusion, vacuum forming,
u. machining of a moulded block or moulding or injection moulding (i.a dependent
on production size),
v. sheet extrusion, bending, welding,
w. spraying a powdery mixture of the two components and a blowing agent onto the
wall; it then foams and hardens,
54 Problems Polymers

x. resin/curing agent blend apply to glass mat (hand-lay-up), curing,

y. injection moulding of pre-form; then blowing-up with bi-axial stretching,
z. sheet extrusion with pentane-injection, vacuum forming.